ONGRID
🌌
18
Ontology·Graph·Decision engine with zero-token ZTC verify
Warecube Korea
Warecube
AI & ML interests
None yet
Recent Activity
liked a Space 1 day ago
FINAL-Bench/ONGRID liked a model 2 days ago
FINAL-Bench/Darwin-27B-ZTC-v2 liked a Space 2 days ago
hotchpotch/S1MB-leaderboardOrganizations
None yet
liked a
Space 1 day ago
liked a
model 2 days ago
liked a
Space 2 days ago
liked 9
datasets 3 days ago
joelniklaus/LEXam-hard
Benchmark • Updated • 518 • 864 • 24
LiquidAI/ifstruct-v1.0
Benchmark • Updated • 2k • 610 • 103
MathArena/aime_2026
Benchmark • Updated • 30 • 31.5k • 86
MathArena/hmmt_feb_2026
Benchmark • Updated • 33 • 18.5k • 31
LEXam-Benchmark/LEXam
Benchmark • Updated • 7.54k • 1.73k • 71
llamaindex/ExtractBench
Benchmark • Updated • 370 • 15.7k • 58
MMMU/MMMU_Pro
Benchmark • Updated • 5.19k • 19.1k • 91
TIGER-Lab/MMLU-Pro
Benchmark • Updated • 12.1k • 223k • 546
Delores-Lin/MDPBench
Benchmark • Updated • 1.64k • 48
upvoted a collection 3 days ago
liked 3
models 3 days ago
upvoted an article 3 days ago
Article
Darwin-27B-ZTC: A Single-Pass Judge and a Quantitative Look at Its Calibration
FINAL-Bench
• • 12 liked a
model 3 days ago
liked a
dataset 3 days ago
reacted to SeaWolf-AI's post with ❤️ 3 days ago
Post
4006
🧠 We just released Darwin-27B-ZTC, a judgment engine that reaches a verdict without generating anything.
Most LLMs answer by generating, decoding one token at a time. Darwin-27B-ZTC takes a different route.
⚙️ How it works
🔹 It makes its call in a single forward pass.
🔹 Zero generated tokens, and no decoding loop.
🔹 That keeps latency and cost far below what a generative model needs.
🎯 What it judges
🔹 It handles several question types: free-form correctness (noul), multiple choice (choice), and scoring (score).
🔹 For each one it hands back a calibrated confidence, not just an answer.
📊 How well calibrated (measured)
🔹 KL 0.204, Brier 0.097, so the confidence it reports lines up with what actually happens.
🔹 0.743 accuracy (zero-shot, general split), across 2,000 judgments with zero errors.
🔹 By type: noul 0.847, choice 0.723, score 0.675.
🔹 None of the benchmark's train split went into it. It is pure zero-shot.
🚀 Where it fits
🔹 Grading at scale, model routing, safety gating, anywhere you want a fast decision without paying for generation.
🏆 It currently sits at #1 on the official typed-decisions leaderboard on Hugging Face (0.743 accuracy, zero-shot).
🔗 Links
Model: FINAL-Bench/Darwin-27B-ZTC
Leaderboard: LocalLLaMA/typed-decisions
Curious to hear what you make of the single-pass, no-generation approach. 🙌
Most LLMs answer by generating, decoding one token at a time. Darwin-27B-ZTC takes a different route.
⚙️ How it works
🔹 It makes its call in a single forward pass.
🔹 Zero generated tokens, and no decoding loop.
🔹 That keeps latency and cost far below what a generative model needs.
🎯 What it judges
🔹 It handles several question types: free-form correctness (noul), multiple choice (choice), and scoring (score).
🔹 For each one it hands back a calibrated confidence, not just an answer.
📊 How well calibrated (measured)
🔹 KL 0.204, Brier 0.097, so the confidence it reports lines up with what actually happens.
🔹 0.743 accuracy (zero-shot, general split), across 2,000 judgments with zero errors.
🔹 By type: noul 0.847, choice 0.723, score 0.675.
🔹 None of the benchmark's train split went into it. It is pure zero-shot.
🚀 Where it fits
🔹 Grading at scale, model routing, safety gating, anywhere you want a fast decision without paying for generation.
🏆 It currently sits at #1 on the official typed-decisions leaderboard on Hugging Face (0.743 accuracy, zero-shot).
🔗 Links
Model: FINAL-Bench/Darwin-27B-ZTC
Leaderboard: LocalLLaMA/typed-decisions
Curious to hear what you make of the single-pass, no-generation approach. 🙌
- 2 replies