Add Levels 4 and 5 (10 configs, 6400 rows), labelled pending held-out confirmation

LEVELS_4_5.md

@@ -0,0 +1,48 @@
+# ModeBench Levels 4 and 5
+
+Ten configs extending the ModeBench release with two harder difficulty levels,
+five domains each, in the same row surface as Levels 1-3: `problem`, `answer`
+(an executable specification carrying its own `verifier`), and
+`answer_mode_count`.
+
+## How these differ from Levels 1-3
+
+Levels 1-3 were published after a held-out confirmation that matched each
+domain to a fixed reference. **Levels 4 and 5 have not cleared that gate.**
+Every config here records, verbatim from its own dataset identity:
+
+```
+status:             frozen_pending_heldout_confirmation
+difficulty_matched: False
+```
+
+They are frozen, and each domain carries its own confirmation record, recipe
+and protocol hashes, and cross-source disjointness checks. They back the
+published base-model scale figure. What they do not yet carry is the held-out
+confirmation and the difficulty match that Levels 1-3 carry, so a result on
+Level 4 or 5 should not be reported as though it rested on the same warrant
+as one on Levels 1-3.
+
+## Splits
+
+| level | domains | train | dev | eval |
+| --- | --- | --- | --- | --- |
+| 4 | 5 | 384 | 128 | 128 |
+| 5 | 5 | 384 | 128 | 128 |
+
+Every split was verified against the row digest frozen in its dataset identity,
+then round-tripped through Parquet and re-read to confirm rows, feature types
+and column order are unchanged.
+
+## Configs
+
+- `level4_countdown` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level4_graph_coloring` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level4_mathir` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level4_pantry_plan` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level4_python_factors` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level5_countdown` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level5_graph_coloring` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level5_mathir` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level5_pantry_plan` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`
+- `level5_python_factors` — status `frozen_pending_heldout_confirmation`, difficulty_matched `False`

README.md

@@ -126,6 +126,86 @@ configs:
     path: data/level3_python_factors/dev.parquet
   - split: eval
     path: data/level3_python_factors/eval.parquet
+- config_name: level4_countdown
+  data_files:
+  - split: dev
+    path: data/level4_countdown/dev.parquet
+  - split: eval
+    path: data/level4_countdown/eval.parquet
+  - split: train
+    path: data/level4_countdown/train.parquet
+- config_name: level4_graph_coloring
+  data_files:
+  - split: dev
+    path: data/level4_graph_coloring/dev.parquet
+  - split: eval
+    path: data/level4_graph_coloring/eval.parquet
+  - split: train
+    path: data/level4_graph_coloring/train.parquet
+- config_name: level4_mathir
+  data_files:
+  - split: dev
+    path: data/level4_mathir/dev.parquet
+  - split: eval
+    path: data/level4_mathir/eval.parquet
+  - split: train
+    path: data/level4_mathir/train.parquet
+- config_name: level4_pantry_plan
+  data_files:
+  - split: dev
+    path: data/level4_pantry_plan/dev.parquet
+  - split: eval
+    path: data/level4_pantry_plan/eval.parquet
+  - split: train
+    path: data/level4_pantry_plan/train.parquet
+- config_name: level4_python_factors
+  data_files:
+  - split: dev
+    path: data/level4_python_factors/dev.parquet
+  - split: eval
+    path: data/level4_python_factors/eval.parquet
+  - split: train
+    path: data/level4_python_factors/train.parquet
+- config_name: level5_countdown
+  data_files:
+  - split: dev
+    path: data/level5_countdown/dev.parquet
+  - split: eval
+    path: data/level5_countdown/eval.parquet
+  - split: train
+    path: data/level5_countdown/train.parquet
+- config_name: level5_graph_coloring
+  data_files:
+  - split: dev
+    path: data/level5_graph_coloring/dev.parquet
+  - split: eval
+    path: data/level5_graph_coloring/eval.parquet
+  - split: train
+    path: data/level5_graph_coloring/train.parquet
+- config_name: level5_mathir
+  data_files:
+  - split: dev
+    path: data/level5_mathir/dev.parquet
+  - split: eval
+    path: data/level5_mathir/eval.parquet
+  - split: train
+    path: data/level5_mathir/train.parquet
+- config_name: level5_pantry_plan
+  data_files:
+  - split: dev
+    path: data/level5_pantry_plan/dev.parquet
+  - split: eval
+    path: data/level5_pantry_plan/eval.parquet
+  - split: train
+    path: data/level5_pantry_plan/train.parquet
+- config_name: level5_python_factors
+  data_files:
+  - split: dev
+    path: data/level5_python_factors/dev.parquet
+  - split: eval
+    path: data/level5_python_factors/eval.parquet
+  - split: train
+    path: data/level5_python_factors/train.parquet
 ---
 
 # ModeBench
@@ -190,3 +270,16 @@ The support-count metadata and reference specifications are for construction, ch
 ## License provenance
 
 The original repository's Apache-2.0 source-code license is preserved in [licenses/source_repository_APACHE_2.0.txt](licenses/source_repository_APACHE_2.0.txt), and copied source copyright notices remain intact. The frozen dataset roots contain no separate dataset-specific license declaration; this export does not invent new dataset license terms. PantryPlan ingredient provenance records USDA FoodData Central Foundation Foods under CC0/public-domain source terms in [pantry_ingredient_source.json](provenance/pantry_ingredient_source.json).
+
+## Levels 4 and 5: admission status
+
+The ten Level 4 and Level 5 configs (`level4_countdown`, `level4_graph_coloring`, `level4_mathir`, `level4_pantry_plan`, `level4_python_factors`, `level5_countdown`, `level5_graph_coloring`, `level5_mathir`, `level5_pantry_plan`, `level5_python_factors`) are frozen and confirmed per domain, and they back the published base-model scale figure.
+
+They have **not** cleared the held-out confirmation that Levels 1-3 cleared before publication, and they are not difficulty-matched against a fixed reference. Each records its own identity verbatim:
+
+```
+status:             frozen_pending_heldout_confirmation
+difficulty_matched: False
+```
+
+A result measured on Level 4 or 5 should not be reported as resting on the same warrant as one measured on Levels 1-3.

data/level4_countdown/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2508b6201c8efc9919583a88314a2c3c668fb05f9480d1748a7386a70983c01c
+size 14481

data/level4_countdown/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9ffe3a874ea6fb39713c7aa9f10c56895415e3297f4d42247bd07c2d92cbee6
+size 14365

data/level4_countdown/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d5e969b27a0552154ad146fe342d3f1bb96b56b406bfdf36ea83e3c3a20b667
+size 22428

data/level4_graph_coloring/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07076d1716e072a2866193c0a8dc9d012f1c5a17888e926bf2fade03bc16b233
+size 18631

data/level4_graph_coloring/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84789a0071d78197e9f961bf4ab02189e168e8b84bd496902630bc098bbeec36
+size 18437

data/level4_graph_coloring/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1806ce59d5550cc33b5c435242035b334d841fa3a60398143c70009b40c2ff9f
+size 28409

data/level4_mathir/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:931966f5bdea1bc49e2ab0ab92cdf815305efadec9ec3671955409699cfc7237
+size 28689

data/level4_mathir/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef8e32741a37d447c5b62b7e9a52f46760731f00184b8df2ed9dfd57f9ed1a85
+size 28765

data/level4_mathir/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b7046ff0b40f53880c014204ebfb94dccd4629403b7e802edcd7d4bfa4a7b8b4
+size 50853

data/level4_pantry_plan/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0cc1be8c282c8e01bd1ee57d1277a4157eb35e85466b5bab0f709cabfc53230d
+size 62373

data/level4_pantry_plan/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c97438a150b84fa558e611cb32ccce0ae57f17e0b3b2c725219f75302c6e981
+size 61747

data/level4_pantry_plan/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34910a2f86e5131c6e791f6e0a5666e996ce6ee09156ee3f9597d2ccfc1ecaf5
+size 119097

data/level4_python_factors/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48ae468c33c36fe6afd04ff333b2acc77fa434df414eb72ae61d8ca8050d2380
+size 25632

data/level4_python_factors/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4e60141fd94d3a038ef4a3af794623bcbe8664306e347676f84ce13213eed1d
+size 25410

data/level4_python_factors/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f09728ab48729f6a84965617100e92333c8c63dd7a607afa5e2cbe199d82b73
+size 43125

data/level5_countdown/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7bc35dca1ce719eafcab008ee08b957d7563c83eb5468bbaca57a37636b1dc1
+size 15398

data/level5_countdown/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11459a06b49bd0dd61eb668e52ae3f2958c2b1aae22dbc1a9514fe41a28fb3ea
+size 15207

data/level5_countdown/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d80d91bc0faeabc825482f8f6fbd44592a5be2f970350895c25eae093e10f795
+size 24557

data/level5_graph_coloring/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db9d1789c353a2af3e5e6532dcaf2db6de28e22dc52171049d3393e82b038c02
+size 22350

data/level5_graph_coloring/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75ca0480b6e7e113ea681d3a1b1145a09fa2c3de7355ff921919ad708851febc
+size 21928

data/level5_graph_coloring/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed2fe860cc971afc285b52536d9d2458431b26b01d5c616d105461beafa06154
+size 33450

data/level5_mathir/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ee260290035a2d5a30f76481f9c42115640629a3a6efa97073841994f21d8c7
+size 34716

data/level5_mathir/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a82293cd9a2245eb664f1f9082eec83d7232a00308e28a47de9b895a0ca8fd04
+size 34531

data/level5_mathir/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea91001c8fff8c230d2bc7d184343324011286fecb27d3e7ca3a94b4cf2f9250
+size 55510

data/level5_pantry_plan/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22d002000a0bd2f6624e274415c8fa0ae291bfd95a257a938729400ab356d5ae
+size 66774

data/level5_pantry_plan/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:474f112bfcb3abf9c3ac41e49c03e2b4f34e6a1d11670a4ad03768a6f37ac763
+size 65530

data/level5_pantry_plan/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b5f3199b5bb074d8f18b84f054d9950bf20e379e1bdabbb66a1f56879d59327
+size 127834

data/level5_python_factors/dev.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e78331a2eb9b1c8b1f71c0c844325778d65bbbc3106fee933f32e9c838c36ff
+size 23612

data/level5_python_factors/eval.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:496f9d6020fcb906a1ee6b6ec9cb868a4d5478c3f492ab0514d3a825a083b521
+size 23551

data/level5_python_factors/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b48efcf989206137696fb1194f5caa22cfb698d06bc343d688e344ad12568e1
+size 41534