AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Incentive-Aligned LLM Summaries

Yanchen Jiang, Zhe Feng, Aranyak Mehta

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Statistical frameworks for LLM evaluation

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
replace majority-claims re-summary with peer-prediction source filtering (TTS)LLM summarization pipeline · NQanswer accuracy: 38.6% -> 72.3%Improved
replace majority-claims re-summary with peer-prediction source filtering (TTS)LLM summarization pipeline · ClashEvalanswer accuracy: 38.4% -> 77.1%Improved
replace initial summary with TTSLLM summarization pipeline · NQprecision: 40.8% -> 76.1%Improved
replace initial summary with TTSLLM summarization pipeline · ClashEvalprecision: 49.3% -> 86.2%Improved
replace initial summary with TTSLLM summarization pipeline · NQF1: 31.9% -> 50.0%Improved
replace initial summary with TTSLLM summarization pipeline · ClashEvalF1: 30.4% -> 56.1%Improved
replace initial summary with TTSLLM summarization pipeline · NQrecall: 26.2% -> 37.2%Improved
replace initial summary with TTSLLM summarization pipeline · ClashEvalrecall: 22.0% -> 31.5%Improved
prompt LLM to use majority claims only instead of plain summaryLLM summarization pipeline · ClashEvalanswer accuracy: 15.6% -> 30.2%Improved
prompt LLM to use majority claims only instead of plain summaryLLM summarization pipeline · NQanswer accuracy: 25.1% -> 27.5%Improved
majority-claims filtering lowers recall relative to initial summaryInitial Synthesis · ClashEvalrecall: 22.0% -> 20.4%Hurt
use output-agreement majority scoring instead of multi-task peer prediction with 4 uninformative sources (ablation)source scoring · ClashEvalavg score of adversarial sourceHurt

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