AI Paper Atlas

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

Semantic Voting: A Self-Evaluation-Free Approach for Efficient LLM Self-Improvement on Unverifiable Open-ended Tasks

Chunyang Jiang, Yonggang Zhang, Yiyang Cai, Chi-Min Chan, Yulong Liu, Mingming Chen, Wei Xue, Yike Guo

ICLR 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 6 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
use SVSI instead of entropy minimizationLlama-8B · WMT24++ de-enBLEU: 0 -> 20.76Improved
use SVSI instead of entropy minimizationLlama-3B · WMT24++ de-enBLEU: 10.52 -> 8.71Hurt
use SVSI instead of self-judgingQwen-7B · CNN/DailyMailROUGE-L: 7.49 -> 20.69Improved
train with SVSI vs base modelQwen-7B · WMT24++ de-enBLEU: 16.42 -> 20.19Improved
train with SVSI vs base modelLlama-8B · WMT24++ de-enBLEU: 20.22 -> 20.76No clear effect
self-judging pseudo-labelsLlama-8B · CNN/DailyMailROUGE-L: 20.68 -> 0.08Hurt
add self-generation clustering filter (ablation)SVSI · Llama-1BperformanceImproved
add self-generation clustering filter (ablation)SVSI · Qwen-1.5BperformanceNo clear effect
add semantic voting on top of clustering-only pairs (ablation)SVSI · translation and summarizationperformanceImproved
flip SV preference pairs (ablation)SVSI · translation and summarizationperformanceHurt
use DeBERTaV3 embeddings instead of SimCSE (ablation)SVSI · WMT24++ avg BLEU, Llama-1BBLEU: 5.13 -> 3.54Hurt
use Mean Shift clustering instead of HDBSCAN (ablation)SVSI · WMT24++ avg BLEU, Llama-1BBLEU: 5.13 -> 3.85Hurt

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