LongCapHarness: Capability-Guided Auto-Harnessing for Long-Context Reasoning
cs.CL
Submitted: 2026-11-07
Updated: 2026-11-07
Terminology
Sources
- HarnessX: A Composable, Adaptive, and Evolvable Agent Harness Foundry
- Automated Design of Agentic Systems
- EntropyLong: Effective Long-Context Training via Predictive Uncertainty
- Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses
- AutoHarness: improving LLM agents by automatically synthesizing a code harness
- TTHE: Test-Time Harness Evolution
- Qwen2.5 Technical Report
- RRSI: Regularized Recursive Self-Improvement of Agent Harnesses
- AFlow: Automating Agentic Workflow Generation
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data