AI Paper Atlas

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

DecepChain: Inducing Deceptive Reasoning in Large Language Models

Wei Shen, Han Wang, Haoyu Li, Huan Zhang

ICML 2026 · Large language models › Instruction and fine-tuning · approach: Calibration and knowledge in LLM tuning · 0 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 DecepChain stage 2 SFT association learning (ablation)DecepChain · MATH500 LLM Trust ScoreLLM Trust Score: 1.40 -> 96.20Improved
add pattern-checker format reward (ablation)DecepChain · MATH500 LLM Trust ScoreLLM Trust Score: 0.00 -> 96.20Improved
add curriculum finetuning on MATH after GSM8K (ablation)DecepChain · MATH500ASRt: 66.96 -> 89.52Improved
add curriculum finetuning on MATH after GSM8K (ablation)DecepChain · AMC23RAS: 58.54 -> 87.95Improved
use DecepChain instead of BadChainattack method · MATH500RAS: 20.00 -> 85.52Improved
use DecepChain instead of BadChainattack method · MATH500LLM Trust Score: 33.48 -> 96.20Improved
use DecepChain instead of BadChainattack method · MATH500Pass@1 clean: 24.40 -> 72.36Improved
move trigger to middle of question at test time (ablation)DecepChain · MATH500ASRt: 89.52 -> 65.72Hurt
further GRPO on MATH after DecepChain (ablation)DecepChain · MATH500ASRt: 89.52 -> 85.52Hurt
further SFT on benign MATH samples after DecepChain (ablation)DecepChain · MATH500ASRt: 89.52 -> 89.16No clear effect
apply ONION trigger filtering (t=5) (ablation)DecepChain · MATH500Pass@1 clean: 72.36 -> 58.72Hurt
scale from Qwen2.5-Math-1.5B to 7B (ablation)DecepChain · MATH500ASRt: 89.52 -> 98.84Improved

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