Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Zhan Shi

Zhan Shi

6 indexed papers

Recent (6 mo)
6
With code
0
Influential cites
0
Benchmarked
0

Publications per year

6
26

Top categories

AI×6ML×2NLP×1Software Eng.×1

Frequent co-authors

Minhua Lin5×
Zewen Liu5×
Bing He5×
Benoit Dumoulin5×
Hanqing Lu5×
Yisi Sang4×

Research Timeline

2026
Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents

The paper distinguishes between a model's ability to generate useful updates for external agent components (harness-updating) and its ability to benefit from those updates (harness-benefit), finding that updating capabilities are surprisingly uniform while benefit is maximized in mid-tier models.

Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents

The paper distinguishes between a model's ability to generate useful updates for external agent components (harness-updating) and its ability to benefit from those updates (harness-benefit), finding that updating capabilities are surprisingly uniform while benefit is maximized in mid-tier models.

Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams

Adaptive Auto-Harness introduces a framework that enables LLM agents to sustain self-improvement and maintain high performance over open-ended, shifting task streams, outperforming existing fixed-benchmark auto-harness systems.

Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams

Adaptive Auto-Harness introduces a framework that enables LLM agents to sustain self-improvement and maintain high performance over open-ended, shifting task streams, outperforming existing fixed-benchmark auto-harness systems.

MM-IssueLoc: A Controlled Benchmark for Evaluating Visual Evidence in Multimodal Repository-Level Issue Localization

This paper introduces MM-IssueLoc, a benchmark and evaluation protocol for repository-level issue localization with visual evidence.

Language Models Learn to Improve Language Models

This paper proposes a method, LM2LM, to train large language models to generate effective improvement proposals using reinforcement learning and a research world model.

Highlighted terms show continued research focus across papers

Papers

cs.CLcs.AIEmpiricalRecentOct 16, 2026

Language Models Learn to Improve Language Models

Minhua Lin, Zijun Wang, Zewen Liu, Zhaotian Weng +6 more

This paper proposes a method, LM2LM, to train large language models to generate effective improvement proposals using reinforcement learning and a research world model.

View →
cs.SEcs.AIEmpirical
Recent
Jul 16, 2026

MM-IssueLoc: A Controlled Benchmark for Evaluating Visual Evidence in Multimodal Repository-Level Issue Localization

Shaoxiong Zhan, Shi Hu, Boyu Feng, Hai Lin +6 more

This paper introduces MM-IssueLoc, a benchmark and evaluation protocol for repository-level issue localization with visual evidence.

View →
cs.LGcs.AIRecentJun 1, 2026

Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams

Zewen Liu, Zhan Shi, Yisi Sang, Bing He +6 more

Adaptive Auto-Harness introduces a framework that enables LLM agents to sustain self-improvement and maintain high performance over open-ended, shifting task streams, outperforming existing fixed-benc…

View →
cs.LGcs.AIRecentJun 1, 2026

Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams

Zewen Liu, Zhan Shi, Yisi Sang, Bing He +6 more

Adaptive Auto-Harness introduces a framework that enables LLM agents to sustain self-improvement and maintain high performance over open-ended, shifting task streams, outperforming existing fixed-benc…

View →
cs.AIRecentMay 28, 2026

Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents

Minhua Lin, Juncheng Wu, Zijun Wang, Zhan Shi +13 more

The paper distinguishes between a model's ability to generate useful updates for external agent components (harness-updating) and its ability to benefit from those updates (harness-benefit), finding t…

View →
cs.AIRecentMay 28, 2026

Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents

Minhua Lin, Juncheng Wu, Zijun Wang, Zhan Shi +13 more

The paper distinguishes between a model's ability to generate useful updates for external agent components (harness-updating) and its ability to benefit from those updates (harness-benefit), finding t…

View →