paper-with-me

홈 › Papers

It's Not the Capability: Harness Sensitivity Is Non-Monotone Across LLM Agent Tiers

2026-05-26 · Yong-eun Cho arxiv

A prevalent assumption in LLM agent deployment holds that more structured harnesses universally improve reliability, and that higher-capability models need proportionally less structural guidance -- together implying a monotone inverse relationship between model capability tier and optimal harness complexity. We test this hypothesis through a controlled 432-run experiment crossing six models across four capability tiers with three harness conditions (light, balanced, strict) on HEAT-24, a 24-task synthetic benchmark with git-based workspace verification. Our results refute the monotone inverse relationship on two fronts. First, for the frontier chat model evaluated (Gemini 2.5 Flash), increased harness verbosity lowers VTSR by 29-38 percentage points -- a harness-complexity paradox. Second, for the frontier reasoning model evaluated (Qwen3.5-122B, extended thinking enabled), strict harness achieves the highest VTSR (91.7%) and the lowest latency, the opposite of the prediction. Within the constrained tier, a 2B model (Gemma4:e2B) matches strong-open-tier stability at 91.7% across all harnesses. Because each tier is represented by a single model in this study, these results should be interpreted as model-specific observations; harness sensitivity appears non-monotone across the models evaluated, and depends critically on model type (chat vs. reasoning). We introduce a six-label failure taxonomy showing that format_violation dominates capable-model failures while wrong_file dominates low-capability failures, and we derive practical tier-aware harness selection guidelines.

📄 PDF Abstract BibTeX arXiv:2605.26731

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Evo-Bench: Can Language Models Improve Agent Harness?

2026-08-10 · Lisheng Huang, Chen Yang, Hao Zhou, Huatong Song 외 hf

Large Language Models (LLMs) have driven rapid progress in autonomous agents, yet standard evaluations remain confined to static task solving. An emerging frontier is harness evolution---the agent's capacity to autonomou…

JIT-Agent: Scaling Harness Intelligence via Just-in-Time Harness Evolution

2026-08-26 · Guibin Zhang, Leo Lu, Fangzhou Xie, Kang Zhu 외 arxiv

Agent capability is not determined by the model alone. The agent harness, encompassing memory management, planning strategy, action protocol, and tool/skill orchestration, can dominate the contribution of the underlying …

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

2026-05-28 · Minhua Lin, Juncheng Wu, Zijun Wang, Zhan Shi 외 arxiv

LLM agents are increasingly deployed as systems built around editable external harnesses, including prompts, skills, memories and tools, that shape task execution without changing model parameters. Harness self-evolution…

Instruction Following

HarnessOpt-Bench: Evaluating LLMs at Harness Optimization

2026-08-06 · Varun Ursekar, Apaar Shanker, Yash Maurya, Shehab Yasser 외 hf

As LLMs are increasingly deployed within agentic systems, their capabilities depend not only on the model weights but also on the harness: the prompts, tools, control flow, memory, and orchestration code surrounding them…

VeRO: A Harness for Agents to Optimize Agents

2026-02-25 · Varun Ursekar, Apaar Shanker, Veronica Chatrath, Yuan Xue 외 arxiv

An important emerging application of coding agents is agent harness optimization: the iterative improvement of a target agent by editing and evaluating its code. Despite its relevance, the community lacks a systematic un…