AI Paper Atlas

Research map / Large language models / Alignment and preferences

Teach2Eval: An Interaction-Driven LLMs Evaluation Method via Teaching Effectiveness

Yuhang Zhou, Xutian Chen, Yixin Cao, Yuchen Ni, Yu He, Siyu Tian, Xiang Liu, Yunwen Chen, Guangnan Ye, Xipeng Qiu, Hongfeng Chai

ICLR 2026 · Large language models › Alignment and preferences · approach: LLM-as-judge and feedback 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
evaluate via teaching weaker students instead of direct answer scoringLLM evaluation · Chatbot ArenaSpearman correlation: 0.734 -> 0.944Improved
evaluate via teaching weaker students instead of direct answer scoringLLM evaluation · LiveBenchSpearman correlation: 0.861 -> 0.975Improved
evaluate via teaching instead of direct scoringLLM evaluation · Chatbot ArenaKendall's Tau: 0.558 -> 0.853Improved
evaluate via teaching instead of direct scoringLLM evaluation · LiveBenchKendall's Tau: 0.695 -> 0.886Improved
remove one of four student models at random (ablation)Teach2Eval · Chatbot ArenaSpearman correlation: 0.926-0.936No clear effect
increase cumulative sample size, Teach2Eval vs direct (ablation)Teach2Eval · LiveBenchKendall's Tau: 0.83 vs near 0.64Improved
more guidance rounds (up to six) (ablation)student models · guidance roundsstudent gainNo clear effect
teacher reveals the answer vs not (ablation)Teach2Eval ranking · answer revelation subsetsSpearman vs Teach2Eval: 0.947 -> 0.921No clear effect
distill GPT-4o answers into models (contamination-like fine-tuning)six LLMs · AA vs CAApplication and Comprehensive AbilityMixed

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