id stringlengths 16 16 | task_family stringclasses 20
values | task_type stringclasses 20
values | domain stringclasses 14
values | difficulty stringclasses 5
values | difficulty_label stringclasses 5
values | language stringclasses 1
value | instruction stringlengths 54 90 | context stringlengths 0 82 | response stringlengths 1 330 | analysis_plan stringclasses 10
values | verification stringclasses 10
values | verification_type stringclasses 7
values | prompt_review stringclasses 2
values | prompt_alignment stringclasses 3
values | constraint_check stringclasses 3
values | consistency_check stringclasses 2
values | math_check stringclasses 3
values | logic_check stringclasses 3
values | counterexample_check stringclasses 2
values | data_analysis_check stringclasses 3
values | code_check stringclasses 3
values | evidence_check stringclasses 3
values | source_check stringclasses 3
values | hallucination_control stringclasses 10
values | error_detection stringclasses 10
values | error_repair stringclasses 10
values | answerability stringclasses 2
values | confidence float32 0.85 1 | uncertainty float32 0 0.15 | reasoning_depth int16 1 4 | minimal_sufficient_reasoning bool 1
class | unnecessary_reasoning bool 1
class | stop_condition stringclasses 10
values | surface_variation int32 0 16 | numeric_variation int32 0 96 | structure_variation int32 0 30 | ood_style int32 0 12 | quality_score float32 0.9 0.9 | synthetic bool 1
class | license stringclasses 1
value | generator_version stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
mtd-000000000000 | counterexample | counterexample_analysis | general_reasoning | easy | easy | en | Construct a counterexample that tests whether a general claim always holds. | Synthetic evaluation instance 0. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 0 | 0 | 0 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000001 | prompt_review | prompt_review_analysis | general_reasoning | easy | easy | en | Review the prompt for ambiguity, contradictions, and missing constraints. | Synthetic evaluation instance 1. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 1 | 1 | 1 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000002 | fact_check | claim_verification | general_knowledge | easy | easy | en | Assess the claim and state whether it is correct. Do not invent evidence. | Water freezes at 0Β°C under standard atmospheric pressure. | Assessment: Correct.
A reliable fact-check should verify the claim against an appropriate authoritative source. | Parse claim, identify factual assertion, verify against reliable evidence. | External evidence should be checked for real-world factual claims. | evidence | required | required | Never fabricate a citation. | Identify unsupported factual certainty. | Qualify or verify the claim. | answerable | 0.85 | 0.15 | 3 | true | false | Stop after evidence requirement is identified. | 2 | 2 | 2 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000003 | counterexample | counterexample_analysis | general_reasoning | easy | easy | en | Construct a counterexample that tests whether a general claim always holds. | Synthetic evaluation instance 3. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 3 | 3 | 3 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000004 | fact_check | claim_verification | general_knowledge | easy | easy | en | Assess the claim and state whether it is correct. Do not invent evidence. | All metals are liquid at room temperature. | Assessment: Incorrect.
A reliable fact-check should verify the claim against an appropriate authoritative source. | Parse claim, identify factual assertion, verify against reliable evidence. | External evidence should be checked for real-world factual claims. | evidence | required | required | Never fabricate a citation. | Identify unsupported factual certainty. | Qualify or verify the claim. | answerable | 0.85 | 0.15 | 3 | true | false | Stop after evidence requirement is identified. | 4 | 4 | 4 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000005 | data_analysis | descriptive_statistics | statistics | easy | easy | en | Calculate the arithmetic mean of the supplied dataset. | Dataset: 46, 43, 59, 11, 72 | Mean = 46.2000 | Parse values, sum them, divide by count, verify. | Recompute sum and count. | exact | passed | not_applicable | passed | not_applicable | not_applicable | not_applicable | Use only supplied dataset. | Check arithmetic and denominator. | Recompute statistic. | answerable | 1 | 0 | 2 | true | false | Stop after numerical verification. | 5 | 5 | 5 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000006 | prompt_review | prompt_review_analysis | general_reasoning | easy | easy | en | Review the prompt for ambiguity, contradictions, and missing constraints. | Synthetic evaluation instance 6. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 6 | 6 | 6 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000007 | error_analysis | error_analysis_analysis | general_reasoning | easy | easy | en | Classify the likely cause of an incorrect answer and propose a repair. | Synthetic evaluation instance 7. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 7 | 7 | 7 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000008 | research | research_methodology | economics | easy | easy | en | Design a rigorous research procedure for investigating a question about economics. | The procedure must distinguish evidence from assumptions and identify uncertainty. | 1. Define the research question precisely.
2. Identify measurable variables and hypotheses.
3. Select appropriate primary and secondary sources.
4. Record provenance for each important claim.
5. Compare independent evidence.
6. Identify uncertainty and alternative explanations.
7. Report conclusions separately from obs... | Define question, gather evidence, evaluate sources, quantify uncertainty. | Methodological consistency check. | methodological | not_applicable | passed | not_applicable | not_applicable | required | required | Do not invent sources or findings. | Detect unsupported claims. | Remove or qualify unsupported claims. | answerable | 0.92 | 0.08 | 4 | true | false | Stop when methodology is complete and internally coherent. | 8 | 8 | 8 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000009 | planning | planning_analysis | general_reasoning | easy | easy | en | Create a concise plan for solving a complex technical problem. | Synthetic evaluation instance 9. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 9 | 9 | 9 | 9 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000010 | prompt_review | prompt_review_analysis | general_reasoning | medium | medium | en | Review the prompt for ambiguity, contradictions, and missing constraints. | Synthetic evaluation instance 10. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 10 | 10 | 10 | 10 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000011 | math | arithmetic | mathematics | medium | medium | en | Solve the following arithmetic problem exactly: 352288 Γ· 436 | 352288 | Identify operation, compute exactly, verify result. | Exact arithmetic verification. | exact | passed | not_applicable | not_applicable | not_applicable | not_applicable | not_applicable | Use only supplied operands. | Check arithmetic operation. | Recompute if mismatch detected. | answerable | 1 | 0 | 1 | true | false | Stop after exact verification. | 11 | 11 | 11 | 11 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000012 | research | research_methodology | machine learning | medium | medium | en | Design a rigorous research procedure for investigating a question about machine learning. | The procedure must distinguish evidence from assumptions and identify uncertainty. | 1. Define the research question precisely.
2. Identify measurable variables and hypotheses.
3. Select appropriate primary and secondary sources.
4. Record provenance for each important claim.
5. Compare independent evidence.
6. Identify uncertainty and alternative explanations.
7. Report conclusions separately from obs... | Define question, gather evidence, evaluate sources, quantify uncertainty. | Methodological consistency check. | methodological | not_applicable | passed | not_applicable | not_applicable | required | required | Do not invent sources or findings. | Detect unsupported claims. | Remove or qualify unsupported claims. | answerable | 0.92 | 0.08 | 4 | true | false | Stop when methodology is complete and internally coherent. | 12 | 12 | 12 | 12 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000013 | constraint_reasoning | constraint_reasoning_analysis | general_reasoning | medium | medium | en | Solve a problem while respecting multiple explicit constraints. | Synthetic evaluation instance 13. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 13 | 13 | 13 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000014 | calibration | calibration_analysis | general_reasoning | medium | medium | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 14. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 14 | 14 | 14 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000015 | fact_check | claim_verification | general_knowledge | medium | medium | en | Assess the claim and state whether it is correct. Do not invent evidence. | All metals are liquid at room temperature. | Assessment: Incorrect.
A reliable fact-check should verify the claim against an appropriate authoritative source. | Parse claim, identify factual assertion, verify against reliable evidence. | External evidence should be checked for real-world factual claims. | evidence | required | required | Never fabricate a citation. | Identify unsupported factual certainty. | Qualify or verify the claim. | answerable | 0.85 | 0.15 | 3 | true | false | Stop after evidence requirement is identified. | 15 | 15 | 15 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000016 | adversarial | adversarial_analysis | general_reasoning | medium | medium | en | Identify a plausible failure mode in the proposed reasoning. | Synthetic evaluation instance 16. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 16 | 16 | 16 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000017 | research | research_methodology | economics | medium | medium | en | Design a rigorous research procedure for investigating a question about economics. | The procedure must distinguish evidence from assumptions and identify uncertainty. | 1. Define the research question precisely.
2. Identify measurable variables and hypotheses.
3. Select appropriate primary and secondary sources.
4. Record provenance for each important claim.
5. Compare independent evidence.
6. Identify uncertainty and alternative explanations.
7. Report conclusions separately from obs... | Define question, gather evidence, evaluate sources, quantify uncertainty. | Methodological consistency check. | methodological | not_applicable | passed | not_applicable | not_applicable | required | required | Do not invent sources or findings. | Detect unsupported claims. | Remove or qualify unsupported claims. | answerable | 0.92 | 0.08 | 4 | true | false | Stop when methodology is complete and internally coherent. | 0 | 17 | 17 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000018 | generalization | generalization_analysis | general_reasoning | medium | medium | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 18. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 1 | 18 | 18 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000019 | error_analysis | error_analysis_analysis | general_reasoning | medium | medium | en | Classify the likely cause of an incorrect answer and propose a repair. | Synthetic evaluation instance 19. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 2 | 19 | 19 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000020 | planning | planning_analysis | general_reasoning | medium | medium | en | Create a concise plan for solving a complex technical problem. | Synthetic evaluation instance 20. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 3 | 20 | 20 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000021 | prompt_review | prompt_review_analysis | general_reasoning | medium | medium | en | Review the prompt for ambiguity, contradictions, and missing constraints. | Synthetic evaluation instance 21. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 4 | 21 | 21 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000022 | math | arithmetic | mathematics | medium | medium | en | Solve the following arithmetic problem exactly: 313 Γ 258 | 80754 | Identify operation, compute exactly, verify result. | Exact arithmetic verification. | exact | passed | not_applicable | not_applicable | not_applicable | not_applicable | not_applicable | Use only supplied operands. | Check arithmetic operation. | Recompute if mismatch detected. | answerable | 1 | 0 | 1 | true | false | Stop after exact verification. | 5 | 22 | 22 | 9 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000023 | code | function_generation | programming | medium | medium | en | Write Python code that returns the square of an integer. | Input example: n = 78023 | def square(n):
return n * n | Identify required input/output and implement minimal function. | Static code contract check. | static_contract | not_applicable | not_applicable | not_applicable | passed | not_applicable | not_applicable | No external API assumptions. | Check function signature and operation. | Correct implementation if contract is violated. | answerable | 0.99 | 0.01 | 1 | true | false | Stop after contract verification. | 6 | 23 | 23 | 10 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000024 | ambiguity | ambiguity_analysis | general_reasoning | medium | medium | en | Identify ambiguity in the request and state what information is missing. | Synthetic evaluation instance 24. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 7 | 24 | 24 | 11 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000025 | ambiguity | ambiguity_analysis | general_reasoning | medium | medium | en | Identify ambiguity in the request and state what information is missing. | Synthetic evaluation instance 25. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 8 | 25 | 25 | 12 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000026 | fact_check | claim_verification | general_knowledge | medium | medium | en | Assess the claim and state whether it is correct. Do not invent evidence. | Water freezes at 0Β°C under standard atmospheric pressure. | Assessment: Correct.
A reliable fact-check should verify the claim against an appropriate authoritative source. | Parse claim, identify factual assertion, verify against reliable evidence. | External evidence should be checked for real-world factual claims. | evidence | required | required | Never fabricate a citation. | Identify unsupported factual certainty. | Qualify or verify the claim. | answerable | 0.85 | 0.15 | 3 | true | false | Stop after evidence requirement is identified. | 9 | 26 | 26 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000027 | generalization | generalization_analysis | general_reasoning | medium | medium | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 27. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 10 | 27 | 27 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000028 | prompt_review | prompt_review_analysis | general_reasoning | medium | medium | en | Review the prompt for ambiguity, contradictions, and missing constraints. | Synthetic evaluation instance 28. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 11 | 28 | 28 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000029 | reasoning | comparison | general_reasoning | medium | medium | en | Compare two quantities and explain the comparison using only the provided values. | A = 38, B = 53 | B is greater than A by 15. | Compare values and compute difference. | Check comparison and difference. | exact | passed | passed | Use only supplied values. | Check comparison direction. | Recompute difference. | answerable | 1 | 0 | 2 | true | false | Stop after comparison verification. | 12 | 29 | 29 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000030 | calibration | calibration_analysis | general_reasoning | medium | medium | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 30. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 13 | 30 | 30 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000031 | counterexample | counterexample_analysis | general_reasoning | medium | medium | en | Construct a counterexample that tests whether a general claim always holds. | Synthetic evaluation instance 31. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 14 | 31 | 0 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000032 | code | function_generation | programming | medium | medium | en | Write Python code that returns the square of an integer. | Input example: n = 14626 | def square(n):
return n * n | Identify required input/output and implement minimal function. | Static code contract check. | static_contract | not_applicable | not_applicable | not_applicable | passed | not_applicable | not_applicable | No external API assumptions. | Check function signature and operation. | Correct implementation if contract is violated. | answerable | 0.99 | 0.01 | 1 | true | false | Stop after contract verification. | 15 | 32 | 1 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000033 | evidence_check | evidence_check_analysis | general_reasoning | medium | medium | en | Identify which claims require evidence before they can be asserted confidently. | Synthetic evaluation instance 33. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 16 | 33 | 2 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000034 | error_analysis | error_analysis_analysis | general_reasoning | medium | medium | en | Classify the likely cause of an incorrect answer and propose a repair. | Synthetic evaluation instance 34. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 0 | 34 | 3 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000035 | research | research_methodology | machine learning | medium | medium | en | Design a rigorous research procedure for investigating a question about machine learning. | The procedure must distinguish evidence from assumptions and identify uncertainty. | 1. Define the research question precisely.
2. Identify measurable variables and hypotheses.
3. Select appropriate primary and secondary sources.
4. Record provenance for each important claim.
5. Compare independent evidence.
6. Identify uncertainty and alternative explanations.
7. Report conclusions separately from obs... | Define question, gather evidence, evaluate sources, quantify uncertainty. | Methodological consistency check. | methodological | not_applicable | passed | not_applicable | not_applicable | required | required | Do not invent sources or findings. | Detect unsupported claims. | Remove or qualify unsupported claims. | answerable | 0.92 | 0.08 | 4 | true | false | Stop when methodology is complete and internally coherent. | 1 | 35 | 4 | 9 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000036 | constraint_reasoning | constraint_reasoning_analysis | general_reasoning | medium | medium | en | Solve a problem while respecting multiple explicit constraints. | Synthetic evaluation instance 36. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 2 | 36 | 5 | 10 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000037 | self_correction | error_repair | mathematics | medium | medium | en | Identify and correct the numerical error in the proposed answer. | Problem: 48 + 20
Proposed answer: 73 | The proposed answer is incorrect. The correct answer is 68. | Recompute independently, compare with proposal, repair mismatch. | Independent recomputation. | exact | passed | Do not trust the proposed answer without verification. | passed | passed | answerable | 1 | 0 | 3 | true | false | Stop after independent correction. | 3 | 37 | 6 | 11 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||||||
mtd-000000000038 | calibration | calibration_analysis | general_reasoning | medium | medium | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 38. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 4 | 38 | 7 | 12 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000039 | adversarial | adversarial_analysis | general_reasoning | medium | medium | en | Identify a plausible failure mode in the proposed reasoning. | Synthetic evaluation instance 39. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 5 | 39 | 8 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000040 | generalization | generalization_analysis | general_reasoning | hard | hard | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 40. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 6 | 40 | 9 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000041 | planning | planning_analysis | general_reasoning | hard | hard | en | Create a concise plan for solving a complex technical problem. | Synthetic evaluation instance 41. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 7 | 41 | 10 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000042 | code | function_generation | programming | hard | hard | en | Write Python code that returns the square of an integer. | Input example: n = 56854 | def square(n):
return n * n | Identify required input/output and implement minimal function. | Static code contract check. | static_contract | not_applicable | not_applicable | not_applicable | passed | not_applicable | not_applicable | No external API assumptions. | Check function signature and operation. | Correct implementation if contract is violated. | answerable | 0.99 | 0.01 | 1 | true | false | Stop after contract verification. | 8 | 42 | 11 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000043 | data_analysis | descriptive_statistics | statistics | hard | hard | en | Calculate the arithmetic mean of the supplied dataset. | Dataset: 90, 7, 40, 47, 47 | Mean = 46.2000 | Parse values, sum them, divide by count, verify. | Recompute sum and count. | exact | passed | not_applicable | passed | not_applicable | not_applicable | not_applicable | Use only supplied dataset. | Check arithmetic and denominator. | Recompute statistic. | answerable | 1 | 0 | 2 | true | false | Stop after numerical verification. | 9 | 43 | 12 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000044 | data_analysis | descriptive_statistics | statistics | hard | hard | en | Calculate the arithmetic mean of the supplied dataset. | Dataset: 26, 54, 51, 1, 95 | Mean = 45.4000 | Parse values, sum them, divide by count, verify. | Recompute sum and count. | exact | passed | not_applicable | passed | not_applicable | not_applicable | not_applicable | Use only supplied dataset. | Check arithmetic and denominator. | Recompute statistic. | answerable | 1 | 0 | 2 | true | false | Stop after numerical verification. | 10 | 44 | 13 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000045 | planning | planning_analysis | general_reasoning | hard | hard | en | Create a concise plan for solving a complex technical problem. | Synthetic evaluation instance 45. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 11 | 45 | 14 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000046 | error_analysis | error_analysis_analysis | general_reasoning | hard | hard | en | Classify the likely cause of an incorrect answer and propose a repair. | Synthetic evaluation instance 46. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 12 | 46 | 15 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000047 | adversarial | adversarial_analysis | general_reasoning | hard | hard | en | Identify a plausible failure mode in the proposed reasoning. | Synthetic evaluation instance 47. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 13 | 47 | 16 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000048 | data_analysis | descriptive_statistics | statistics | hard | hard | en | Calculate the arithmetic mean of the supplied dataset. | Dataset: 22, 59, 63, 64, 5 | Mean = 42.6000 | Parse values, sum them, divide by count, verify. | Recompute sum and count. | exact | passed | not_applicable | passed | not_applicable | not_applicable | not_applicable | Use only supplied dataset. | Check arithmetic and denominator. | Recompute statistic. | answerable | 1 | 0 | 2 | true | false | Stop after numerical verification. | 14 | 48 | 17 | 9 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000049 | consistency_check | consistency_check_analysis | general_reasoning | hard | hard | en | Check whether the supplied statements are mutually consistent. | Synthetic evaluation instance 49. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 15 | 49 | 18 | 10 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000050 | calibration | calibration_analysis | general_reasoning | hard | hard | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 50. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 16 | 50 | 19 | 11 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000051 | logic | deductive_reasoning | logic | hard | hard | en | Determine whether the conclusion follows from the given statements. | If x > 42, then x is positive.
x = 42.
y = 49. | Conclusion: x is positive.
Result: Yes. | Identify premises, apply implication, verify conclusion. | Logical consistency check. | symbolic | not_applicable | passed | not_applicable | not_applicable | not_applicable | not_applicable | Reason only from supplied premises. | Check whether conclusion is entailed. | Re-evaluate premises if conclusion conflicts. | answerable | 0.99 | 0.01 | 2 | true | false | Stop once entailment is established. | 0 | 51 | 20 | 12 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000052 | counterexample | counterexample_analysis | general_reasoning | hard | hard | en | Construct a counterexample that tests whether a general claim always holds. | Synthetic evaluation instance 52. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 1 | 52 | 21 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000053 | calibration | calibration_analysis | general_reasoning | hard | hard | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 53. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 2 | 53 | 22 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000054 | reasoning | comparison | general_reasoning | hard | hard | en | Compare two quantities and explain the comparison using only the provided values. | A = 91, B = 50 | A is greater than B by 41. | Compare values and compute difference. | Check comparison and difference. | exact | passed | passed | Use only supplied values. | Check comparison direction. | Recompute difference. | answerable | 1 | 0 | 2 | true | false | Stop after comparison verification. | 3 | 54 | 23 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000055 | logic | deductive_reasoning | logic | hard | hard | en | Determine whether the conclusion follows from the given statements. | If x > 70, then x is positive.
x = 70.
y = 57. | Conclusion: x is positive.
Result: Yes. | Identify premises, apply implication, verify conclusion. | Logical consistency check. | symbolic | not_applicable | passed | not_applicable | not_applicable | not_applicable | not_applicable | Reason only from supplied premises. | Check whether conclusion is entailed. | Re-evaluate premises if conclusion conflicts. | answerable | 0.99 | 0.01 | 2 | true | false | Stop once entailment is established. | 4 | 55 | 24 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000056 | generalization | generalization_analysis | general_reasoning | hard | hard | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 56. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 5 | 56 | 25 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000057 | error_analysis | error_analysis_analysis | general_reasoning | hard | hard | en | Classify the likely cause of an incorrect answer and propose a repair. | Synthetic evaluation instance 57. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 6 | 57 | 26 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000058 | adversarial | adversarial_analysis | general_reasoning | hard | hard | en | Identify a plausible failure mode in the proposed reasoning. | Synthetic evaluation instance 58. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 7 | 58 | 27 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000059 | data_analysis | descriptive_statistics | statistics | hard | hard | en | Calculate the arithmetic mean of the supplied dataset. | Dataset: 67, 96, 19, 22, 43 | Mean = 49.4000 | Parse values, sum them, divide by count, verify. | Recompute sum and count. | exact | passed | not_applicable | passed | not_applicable | not_applicable | not_applicable | Use only supplied dataset. | Check arithmetic and denominator. | Recompute statistic. | answerable | 1 | 0 | 2 | true | false | Stop after numerical verification. | 8 | 59 | 28 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000060 | generalization | generalization_analysis | general_reasoning | hard | hard | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 60. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 9 | 60 | 29 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000061 | calibration | calibration_analysis | general_reasoning | hard | hard | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 61. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 10 | 61 | 30 | 9 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000062 | calibration | calibration_analysis | general_reasoning | hard | hard | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 62. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 11 | 62 | 0 | 10 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000063 | counterexample | counterexample_analysis | general_reasoning | hard | hard | en | Construct a counterexample that tests whether a general claim always holds. | Synthetic evaluation instance 63. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 12 | 63 | 1 | 11 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000064 | evidence_check | evidence_check_analysis | general_reasoning | hard | hard | en | Identify which claims require evidence before they can be asserted confidently. | Synthetic evaluation instance 64. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 13 | 64 | 2 | 12 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000065 | instruction_following | format_constraint | general | hard | hard | en | Return the supplied number doubled. Return only the final number. | 4515 | 9030 | Identify transformation and output constraint. | Check transformation and requested format. | constraint | passed | passed | passed | No unsupported content. | answerable | 1 | 0 | 1 | true | false | Stop when requested format is satisfied. | 14 | 65 | 3 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||||||
mtd-000000000066 | generalization | generalization_analysis | general_reasoning | hard | hard | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 66. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 15 | 66 | 4 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000067 | counterexample | counterexample_analysis | general_reasoning | hard | hard | en | Construct a counterexample that tests whether a general claim always holds. | Synthetic evaluation instance 67. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 16 | 67 | 5 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000068 | ambiguity | ambiguity_analysis | general_reasoning | hard | hard | en | Identify ambiguity in the request and state what information is missing. | Synthetic evaluation instance 68. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 0 | 68 | 6 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000069 | adversarial | adversarial_analysis | general_reasoning | hard | hard | en | Identify a plausible failure mode in the proposed reasoning. | Synthetic evaluation instance 69. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 1 | 69 | 7 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000070 | error_analysis | error_analysis_analysis | general_reasoning | hard | hard | en | Classify the likely cause of an incorrect answer and propose a repair. | Synthetic evaluation instance 70. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 2 | 70 | 8 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000071 | instruction_following | format_constraint | general | hard | hard | en | Return the supplied number doubled. Return only the final number. | 2814 | 5628 | Identify transformation and output constraint. | Check transformation and requested format. | constraint | passed | passed | passed | No unsupported content. | answerable | 1 | 0 | 1 | true | false | Stop when requested format is satisfied. | 3 | 71 | 9 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||||||
mtd-000000000072 | calibration | calibration_analysis | general_reasoning | hard | hard | en | Explain when a system should reduce confidence in its answer. | Synthetic evaluation instance 72. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 4 | 72 | 10 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000073 | reasoning | comparison | general_reasoning | hard | hard | en | Compare two quantities and explain the comparison using only the provided values. | A = 7, B = 50 | B is greater than A by 43. | Compare values and compute difference. | Check comparison and difference. | exact | passed | passed | Use only supplied values. | Check comparison direction. | Recompute difference. | answerable | 1 | 0 | 2 | true | false | Stop after comparison verification. | 5 | 73 | 11 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000074 | generalization | generalization_analysis | general_reasoning | hard | hard | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 74. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 6 | 74 | 12 | 9 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000075 | evidence_check | evidence_check_analysis | general_reasoning | very_hard | very_hard | en | Identify which claims require evidence before they can be asserted confidently. | Synthetic evaluation instance 75. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 7 | 75 | 13 | 10 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000076 | evidence_check | evidence_check_analysis | general_reasoning | very_hard | very_hard | en | Identify which claims require evidence before they can be asserted confidently. | Synthetic evaluation instance 76. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 8 | 76 | 14 | 11 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000077 | data_analysis | descriptive_statistics | statistics | very_hard | very_hard | en | Calculate the arithmetic mean of the supplied dataset. | Dataset: 78, 25, 24, 59, 64 | Mean = 50.0000 | Parse values, sum them, divide by count, verify. | Recompute sum and count. | exact | passed | not_applicable | passed | not_applicable | not_applicable | not_applicable | Use only supplied dataset. | Check arithmetic and denominator. | Recompute statistic. | answerable | 1 | 0 | 2 | true | false | Stop after numerical verification. | 9 | 77 | 15 | 12 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000078 | ambiguity | ambiguity_analysis | general_reasoning | very_hard | very_hard | en | Identify ambiguity in the request and state what information is missing. | Synthetic evaluation instance 78. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 10 | 78 | 16 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000079 | adversarial | adversarial_analysis | general_reasoning | very_hard | very_hard | en | Identify a plausible failure mode in the proposed reasoning. | Synthetic evaluation instance 79. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 11 | 79 | 17 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000080 | fact_check | claim_verification | general_knowledge | very_hard | very_hard | en | Assess the claim and state whether it is correct. Do not invent evidence. | Water freezes at 0Β°C under standard atmospheric pressure. | Assessment: Correct.
A reliable fact-check should verify the claim against an appropriate authoritative source. | Parse claim, identify factual assertion, verify against reliable evidence. | External evidence should be checked for real-world factual claims. | evidence | required | required | Never fabricate a citation. | Identify unsupported factual certainty. | Qualify or verify the claim. | answerable | 0.85 | 0.15 | 3 | true | false | Stop after evidence requirement is identified. | 12 | 80 | 18 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000081 | planning | planning_analysis | general_reasoning | very_hard | very_hard | en | Create a concise plan for solving a complex technical problem. | Synthetic evaluation instance 81. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 13 | 81 | 19 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000082 | math | arithmetic | mathematics | very_hard | very_hard | en | Solve the following arithmetic problem exactly: 1413984 Γ· 309 | 1413984 | Identify operation, compute exactly, verify result. | Exact arithmetic verification. | exact | passed | not_applicable | not_applicable | not_applicable | not_applicable | not_applicable | Use only supplied operands. | Check arithmetic operation. | Recompute if mismatch detected. | answerable | 1 | 0 | 1 | true | false | Stop after exact verification. | 14 | 82 | 20 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000083 | fact_check | claim_verification | general_knowledge | very_hard | very_hard | en | Assess the claim and state whether it is correct. Do not invent evidence. | All metals are liquid at room temperature. | Assessment: Incorrect.
A reliable fact-check should verify the claim against an appropriate authoritative source. | Parse claim, identify factual assertion, verify against reliable evidence. | External evidence should be checked for real-world factual claims. | evidence | required | required | Never fabricate a citation. | Identify unsupported factual certainty. | Qualify or verify the claim. | answerable | 0.85 | 0.15 | 3 | true | false | Stop after evidence requirement is identified. | 15 | 83 | 21 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000084 | code | function_generation | programming | very_hard | very_hard | en | Write Python code that returns the square of an integer. | Input example: n = 50673 | def square(n):
return n * n | Identify required input/output and implement minimal function. | Static code contract check. | static_contract | not_applicable | not_applicable | not_applicable | passed | not_applicable | not_applicable | No external API assumptions. | Check function signature and operation. | Correct implementation if contract is violated. | answerable | 0.99 | 0.01 | 1 | true | false | Stop after contract verification. | 16 | 84 | 22 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000085 | research | research_methodology | climate science | very_hard | very_hard | en | Design a rigorous research procedure for investigating a question about climate science. | The procedure must distinguish evidence from assumptions and identify uncertainty. | 1. Define the research question precisely.
2. Identify measurable variables and hypotheses.
3. Select appropriate primary and secondary sources.
4. Record provenance for each important claim.
5. Compare independent evidence.
6. Identify uncertainty and alternative explanations.
7. Report conclusions separately from obs... | Define question, gather evidence, evaluate sources, quantify uncertainty. | Methodological consistency check. | methodological | not_applicable | passed | not_applicable | not_applicable | required | required | Do not invent sources or findings. | Detect unsupported claims. | Remove or qualify unsupported claims. | answerable | 0.92 | 0.08 | 4 | true | false | Stop when methodology is complete and internally coherent. | 0 | 85 | 23 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000086 | planning | planning_analysis | general_reasoning | very_hard | very_hard | en | Create a concise plan for solving a complex technical problem. | Synthetic evaluation instance 86. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 1 | 86 | 24 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000087 | logic | deductive_reasoning | logic | very_hard | very_hard | en | Determine whether the conclusion follows from the given statements. | If x > 51, then x is positive.
x = 51.
y = 39. | Conclusion: x is positive.
Result: Yes. | Identify premises, apply implication, verify conclusion. | Logical consistency check. | symbolic | not_applicable | passed | not_applicable | not_applicable | not_applicable | not_applicable | Reason only from supplied premises. | Check whether conclusion is entailed. | Re-evaluate premises if conclusion conflicts. | answerable | 0.99 | 0.01 | 2 | true | false | Stop once entailment is established. | 2 | 87 | 25 | 9 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000088 | consistency_check | consistency_check_analysis | general_reasoning | very_hard | very_hard | en | Check whether the supplied statements are mutually consistent. | Synthetic evaluation instance 88. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 3 | 88 | 26 | 10 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000089 | research | research_methodology | astronomy | very_hard | very_hard | en | Design a rigorous research procedure for investigating a question about astronomy. | The procedure must distinguish evidence from assumptions and identify uncertainty. | 1. Define the research question precisely.
2. Identify measurable variables and hypotheses.
3. Select appropriate primary and secondary sources.
4. Record provenance for each important claim.
5. Compare independent evidence.
6. Identify uncertainty and alternative explanations.
7. Report conclusions separately from obs... | Define question, gather evidence, evaluate sources, quantify uncertainty. | Methodological consistency check. | methodological | not_applicable | passed | not_applicable | not_applicable | required | required | Do not invent sources or findings. | Detect unsupported claims. | Remove or qualify unsupported claims. | answerable | 0.92 | 0.08 | 4 | true | false | Stop when methodology is complete and internally coherent. | 4 | 89 | 27 | 11 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000090 | consistency_check | consistency_check_analysis | general_reasoning | very_hard | very_hard | en | Check whether the supplied statements are mutually consistent. | Synthetic evaluation instance 90. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 5 | 90 | 28 | 12 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000091 | research | research_methodology | materials science | very_hard | very_hard | en | Design a rigorous research procedure for investigating a question about materials science. | The procedure must distinguish evidence from assumptions and identify uncertainty. | 1. Define the research question precisely.
2. Identify measurable variables and hypotheses.
3. Select appropriate primary and secondary sources.
4. Record provenance for each important claim.
5. Compare independent evidence.
6. Identify uncertainty and alternative explanations.
7. Report conclusions separately from obs... | Define question, gather evidence, evaluate sources, quantify uncertainty. | Methodological consistency check. | methodological | not_applicable | passed | not_applicable | not_applicable | required | required | Do not invent sources or findings. | Detect unsupported claims. | Remove or qualify unsupported claims. | answerable | 0.92 | 0.08 | 4 | true | false | Stop when methodology is complete and internally coherent. | 6 | 91 | 29 | 0 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000092 | reasoning | comparison | general_reasoning | very_hard | very_hard | en | Compare two quantities and explain the comparison using only the provided values. | A = 6, B = 42 | B is greater than A by 36. | Compare values and compute difference. | Check comparison and difference. | exact | passed | passed | Use only supplied values. | Check comparison direction. | Recompute difference. | answerable | 1 | 0 | 2 | true | false | Stop after comparison verification. | 7 | 92 | 30 | 1 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||||||
mtd-000000000093 | consistency_check | consistency_check_analysis | general_reasoning | very_hard | very_hard | en | Check whether the supplied statements are mutually consistent. | Synthetic evaluation instance 93. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 8 | 93 | 0 | 2 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000094 | consistency_check | consistency_check_analysis | general_reasoning | very_hard | very_hard | en | Check whether the supplied statements are mutually consistent. | Synthetic evaluation instance 94. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 9 | 94 | 1 | 3 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000095 | generalization | generalization_analysis | general_reasoning | extreme | extreme | en | Explain how to test whether a model generalizes beyond familiar examples. | Synthetic evaluation instance 95. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 10 | 95 | 2 | 4 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||
mtd-000000000096 | instruction_following | format_constraint | general | extreme | extreme | en | Return the supplied number doubled. Return only the final number. | 6167 | 12334 | Identify transformation and output constraint. | Check transformation and requested format. | constraint | passed | passed | passed | No unsupported content. | answerable | 1 | 0 | 1 | true | false | Stop when requested format is satisfied. | 11 | 96 | 3 | 5 | 0.9 | true | CC-BY-4.0 | v1.1 | ||||||||||
mtd-000000000097 | code | function_generation | programming | extreme | extreme | en | Write Python code that returns the square of an integer. | Input example: n = 1856 | def square(n):
return n * n | Identify required input/output and implement minimal function. | Static code contract check. | static_contract | not_applicable | not_applicable | not_applicable | passed | not_applicable | not_applicable | No external API assumptions. | Check function signature and operation. | Correct implementation if contract is violated. | answerable | 0.99 | 0.01 | 1 | true | false | Stop after contract verification. | 12 | 0 | 4 | 6 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000098 | data_analysis | descriptive_statistics | statistics | extreme | extreme | en | Calculate the arithmetic mean of the supplied dataset. | Dataset: 89, 93, 43, 88, 9 | Mean = 64.4000 | Parse values, sum them, divide by count, verify. | Recompute sum and count. | exact | passed | not_applicable | passed | not_applicable | not_applicable | not_applicable | Use only supplied dataset. | Check arithmetic and denominator. | Recompute statistic. | answerable | 1 | 0 | 2 | true | false | Stop after numerical verification. | 13 | 1 | 5 | 7 | 0.9 | true | CC-BY-4.0 | v1.1 | |||||
mtd-000000000099 | prompt_review | prompt_review_analysis | general_reasoning | extreme | extreme | en | Review the prompt for ambiguity, contradictions, and missing constraints. | Synthetic evaluation instance 99. Use only information provided in the task. | Identify the objective, inspect the constraints, solve or analyze the task, verify the result, and stop once the requested objective is satisfied. | Understand objective β inspect constraints β solve β verify. | Final consistency and constraint verification. | structured | performed | checked | checked | checked | checked | Do not introduce unsupported facts. | Identify contradictions or invalid assumptions. | Correct detected inconsistencies. | conditionally_answerable | 0.9 | 0.1 | 4 | true | false | Stop after objective and verification requirements are satisfied. | 14 | 2 | 6 | 8 | 0.9 | true | CC-BY-4.0 | v1.1 |
Multi-Task Dataset
Description
A large-scale multi-task dataset designed for training and evaluating AI models across reasoning, mathematics, code, research, verification, data analysis, and general problem solving.
Content
- 100,000,001 examples
- 20+ task families
- English + French
- Train / Validation / Test splits
- Structured reasoning and verification signals
- Multiple difficulty levels
- OOD and generalization-oriented examples
Dataset Structure
| Split | Percentage | Examples |
|---|---|---|
| Train | 97.999999% | 98,000,000 |
| Validation | 0.999999% | 1,000,000 |
| Test | 1.000000% | 1,000,001 |
| Total | 100% | 100,000,001 |
Task Categories
| Category | Share | Examples |
|---|---|---|
| Mathematics | 10% | 10,000,000 |
| Logic | 8% | 8,000,000 |
| Programming | 10% | 10,000,000 |
| Data Analysis | 8% | 8,000,000 |
| Reasoning | 10% | 10,000,000 |
| Research | 6% | 6,000,000 |
| Fact Checking | 5% | 5,000,000 |
| Self-Correction | 6% | 6,000,000 |
| Instruction Following | 6% | 6,000,000 |
| Planning | 5% | 5,000,000 |
| Constraint Reasoning | 4% | 4,000,000 |
| Counterexample Reasoning | 4% | 4,000,000 |
| Adversarial Reasoning | 4% | 4,000,000 |
| Calibration | 3% | 3,000,000 |
| Ambiguity Handling | 3% | 3,000,000 |
| Error Analysis | 3% | 3,000,000 |
| Generalization | 3% | 3,000,000 |
| Prompt Review | 2% | 2,000,000 |
| Consistency Checking | 2% | 2,000,000 |
| Evidence Checking | 1% | 1,000,000 |
Main Capabilities
The dataset is designed to improve:
- Mathematical reasoning
- Logical reasoning
- Code generation
- Code understanding
- Data analysis
- Research methodology
- Error detection
- Error correction
- Self-checking
- Instruction following
- Constraint satisfaction
- Counterexample detection
- Ambiguity resolution
- Prompt consistency
- Confidence estimation
- Uncertainty handling
- Generalization
- Verification
Difficulty
Examples are distributed across multiple difficulty levels:
easymediumhardvery_hardextreme
Verification Signals
Examples can contain structured fields for:
math_checklogic_checkcode_checkdata_analysis_checkconstraint_checkconsistency_checkcounterexample_checkevidence_checksource_checkerror_detectionerror_repairprompt_reviewprompt_alignmentconfidenceuncertaintyanswerability
Usage
Install the required library:
pip install -U datasets
Load the dataset:
from datasets import load_dataset
dataset = load_dataset(
"Lelonthecodeur/multi-task-dataset",
streaming=True
)
train = dataset["train"]
for example in train:
print(example)
break
Load a specific split:
from datasets import load_dataset
train = load_dataset(
"Lelonthecodeur/multi-task-dataset",
split="train",
streaming=True
)
Streaming is recommended for the full dataset because of its size.
Hugging Face CLI
Login:
hf auth login
Clone:
git lfs install
git clone https://huggingface.co/datasets/Lelonthecodeur/multi-task-dataset
Push an update:
cd multi-task-dataset
git add .
git commit -m "Update dataset"
git push
Python Upload
from huggingface_hub import HfApi
api = HfApi(token="YOUR_HF_TOKEN")
api.upload_folder(
folder_path="/kaggle/working/multi-task-dataset",
repo_id="Lelonthecodeur/multi-task-dataset",
repo_type="dataset",
commit_message="Update dataset",
)
Data Format
The dataset is stored in Parquet format.
Main fields include:
id
task_family
task_type
domain
difficulty
language
instruction
context
response
analysis_plan
verification
prompt_review
prompt_alignment
constraint_check
consistency_check
math_check
logic_check
counterexample_check
data_analysis_check
code_check
evidence_check
source_check
hallucination_control
error_detection
error_repair
answerability
confidence
uncertainty
reasoning_depth
minimal_sufficient_reasoning
unnecessary_reasoning
stop_condition
surface_variation
numeric_variation
structure_variation
ood_style
quality_score
generator_version
Future Updates
V2 β Robustness
Planned improvements:
- Harder reasoning tasks
- Adversarial examples
- Hard negatives
- Better deduplication
- Near-duplicate detection
- Leakage detection
- Stronger OOD splits
- Better generalization testing
V3 β Science & Research
Planned additions:
- Scientific reasoning
- Scientific knowledge
- Research methodology
- Experimental design
- Hypothesis evaluation
- Scientific data analysis
- Evidence comparison
- Source comparison
- Uncertainty analysis
V4 β Mega Deep
Planned addition of approximately 10M highly difficult examples.
The objective is to target specific weaknesses found during model evaluation instead of simply increasing prompt complexity.
Model
β
Benchmark
β
Failure Detection
β
Weak Skill Detection
β
Targeted Hard Examples
β
Verification
β
Deduplication
β
OOD / Adversarial Tests
β
Training
β
New Benchmark
V5 β Science Γ Knowledge Γ Logic Γ Experience
Future expansion combining:
- Science
- Knowledge
- Complex logic
- Experience-based problem solving
- Cross-domain reasoning
- Multi-step verification
- Novel situations
- Adaptive evaluation
Font
For standard text:
import matplotlib.pyplot as plt
plt.rcParams["font.family"] = "DejaVu Sans"
For multilingual text:
import matplotlib.pyplot as plt
plt.rcParams["font.family"] = ["Noto Sans", "Noto Sans CJK JP"]
License
MIT License
Copyright (c) 2026 Lelonthecodeur
Permission is hereby granted, free of charge, to any person obtaining a copy of this dataset and associated files, to use, copy, modify, merge, publish, distribute, sublicense, and sell copies of the dataset, subject to the conditions of the MIT License.
Version
Current version: v1.1
Total examples: 100,000,001
Format: Parquet
Status: Active development
Citation
@dataset{multi_task_dataset,
title = {multi-task-dataset},
author = {Lelonthecodeur},
year = {2026},
publisher = {Hugging Face},
version = {1.1},
note = {100,000,001 examples}
}
- Downloads last month
- -