研究库 论文知识库

内容库 / 论文

arXiv:2608.31100 · 评测基准

S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?

S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?

类型：arxiv
标识：2608.31100
链接：https://arxiv.org/abs/2608.31100
主分类：evaluation
形态：benchmark
TLDR：Large language models (LLMs) increasingly interact with external environments and accumulate substantial behavioral experience, yet existing agent benchmarks largely evaluate them as fixed policies. It therefore remains unclear whether an agent can actively test its behavior, judge the resulting experience, and use that experience to improve future decisions. We introduce S\textsuperscript{3Gym}, an interactive benchmark for evaluating LLM self-improvement through three coupled capabilities: Self-Testing, Self-Judging, and Self-Improvement. S^3Gym separates permissive exploration from strict h
副分类：agent
待LLM分类：否
来源文件：
/inbox/tom/_candidates/2026-09-03-agent-rag-longcontext-candidates.json

相关 · 按语义匹配

仓库 Dusto-AI/llm-situational-awareness-paper-list

笔记精读与批判 · From 128K to 4M: Efficient Training of Ultra-Long Context LLMs

笔记 Cool Papers · cs.CL (papers.cool) · RSS 摘要

仓库 DSXiangLi/DecryptPrompt

仓库 thuyhuongctu/M-AIDA

仓库 PranayMahendrakar/arxiv-llm-analyzer

arXiv

2608.31100

TLDR

Large language models (LLMs) increasingly interact with external environments and accumulate substantial behavioral experience, yet existing agent benchmarks largely evaluate them as fixed policies. It therefore remains unclear whether an agent can actively test its behavior, judge the resulting experience, and use that experience to improve future decisions. We introduce S\textsuperscript{3Gym}, an interactive benchmark for evaluating LLM self-improvement through three coupled capabilities: Self-Testing, Self-Judging, and Self-Improvement. S^3Gym separates permissive exploration from strict h

相关解读

arXiv ↗ 中文全文