AI Paper Atlas

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

Anchored Supervised Fine-Tuning

He Zhu, Junyou Su, Peng Lai, Ren Ma, Wenjia Zhang, Linyi Yang, Guanhua Chen

ICLR 2026 · Large language models › Instruction and fine-tuning · approach: Fine-tuning for reasoning data

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 forward-KL anchoring to DFT (ablation)DFT · Medical avg, 10k, LLaMA-2-7Baccuracy: 29.19 -> 42.03Improved
add forward-KL anchoring to DFT (ablation)DFT · Math avg, 100k, Qwen2.5-7Baccuracy: 26.04 -> 30.50Improved
add forward-KL anchoring to DFT (ablation)DFT · Math avg, 30k, Qwen2.5-7Baccuracy: 27.66 -> 27.18Hurt
add probability reweighting (DFT) to SFTSFT · Medical avg, 10k, LLaMA-2-7Baccuracy: 33.37 -> 29.19Hurt
add probability reweighting (DFT) to SFTSFT · Math avg, 10k, Qwen2.5-7Baccuracy: 16.73 -> 27.77Improved
add plain KL regularization to SFT (ablation)SFT · Medical avg, 10k, LLaMA-2-7Baccuracy: 33.37 -> 30.95Hurt
add plain KL regularization to SFT (ablation)SFT · Math avg, 10k, Qwen2.5-7Baccuracy: 16.73 -> 16.79No clear effect
use forward KL instead of reverse KL as anchor (ablation)ASFT · MedQA, MMLU, MedMCQA (LLaMA-2-7B)accuracyImproved
use SFT loss as anchor instead of KL, large coefficient (ablation)ASFT-SFT · Math, alpha 0.01 -> 10.0accuracy: 0.275 -> 0.168Hurt
use SFT loss as anchor instead of KL (ablation)ASFT-SFT · Medical, varying alphaaccuracy: fluctuating between 0.340 and 0.370Mixed
initialize DAPO from ASFT instead of SFTDAPO · Medical avg, LLaMA-2-7Baccuracy: 40.24 -> 44.10Improved
use ASFT instead of SFTSFT · Code avg, LLaMA-2-7Bpass@1 (%): 26.4 -> 27.0Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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