AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

Probability-Entropy Calibration: An Elastic Indicator for Adaptive Fine-tuning

Wenhao Yu, Shaohang Wei, Jiahong Liu, Yifan Li, Minda Hu, Aiwei Liu, Hao Zhang, Irwin King

ICML 2026 · Large language models › Instruction and fine-tuning · approach: Fine-tuning for reasoning data · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add the entropy term (calibrate probability by expected rank) (ablation)RankTuner (Qwen2.5-Math-7B) · AIME24 Pass@kPass@kImproved
add the probability term (ablation)RankTuner (Qwen2.5-Math-7B) · AIME24 Pass@kPass@1 / Pass@kMixed
use dynamic exponent K(xi) instead of fixed exponents (ablation)RankTuner (Qwen2.5-Math-7B) · math reasoningPass@kImproved
use RankTuner instead of Alpha Power loss (alpha=0.5)Qwen2.5-Math-7B · AIME24 / OlympiadBenchPass@16Improved
use RankTuner instead of entropy regularization (alpha=0.02)Qwen2.5-Math-7B · AIME24 / OlympiadBenchPass@16Improved
apply RankTuner reweighting on top of base modelQwen2.5-Math-7B · MATH-OAIPass@1: 31.79 -> 68.60Improved
apply RankTuner reweighting on top of base modelQwen3-8B · MATH-OAIPass@1: 65.14 -> 72.38Improved
plain SFT instead of the original modelQwen3-8B · MATH-OAIPass@1: 65.14 -> 54.83Hurt
use RankTuner instead of TALRQwen3-8B · Minerva MathPass@1: 40.46 -> 38.26Hurt
use RankTuner instead of best baselineQwen2.5-Math-7B · AMC23Pass@16: 82.50 -> 82.50No clear effect
use RankTuner instead of DFTQwen2.5-Math-7B · ARC-C and GPQA (OOD)Pass@1: 26.50 -> 53.58 (ARC-C)Improved
use RankTuner instead of entropy- or prob-dominant weightingnoise diagnostic · injected noisesequence noise hit@10%: 77% -> 9%Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy