Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents
cs.AI
Submitted: 2026-05-28
Updated: 2026-11-04
Comments: Accepted by EMNLP 2026 (main)
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- PromptWizard: Task-Aware Prompt Optimization Framework
- gpt-oss-120b & gpt-oss-20b Model Card
- EvoSkill: Automated Skill Discovery for Multi-Agent Systems
- MCP-Atlas: A Large-Scale Benchmark for Tool-Use Competency with Real MCP Servers
- Measuring Massive Multitask Language Understanding
- SEA-Eval: A Benchmark for Evaluating Self-Evolving Agents Beyond Episodic Assessment
- Meta-Harness: End-to-End Optimization of Model Harnesses
- Yunjue Agent Tech Report: A Fully Reproducible, Zero-Start In-Situ Self-Evolving Agent System for Open-Ended Tasks
- SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks
- A Comprehensive Survey on Reinforcement Learning-based Agentic Search: Foundations, Roles, Optimizations, Evaluations, and Applications
- MemMA: Coordinating the Memory Cycle through Multi-Agent Reasoning and In-Situ Self-Evolution
- Graph-of-Skills: Dependency-Aware Structural Retrieval for Massive Agent Skills
- Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces
- Code as Agent Harness
- ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory
- Natural-Language Agent Harnesses
- LLaMA: Open and Efficient Foundation Language Models
- Voyager: An Open-Ended Embodied Agent with Large Language Models
- Evaluation of LLMs for mathematical problem solving
- Agent Workflow Memory
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents