{"slug": "clueweaver-reward-guided-dual-agent-evidence-reasoning-for-compact-llms-on-long", "title": "ClueWeaver: Reward-Guided Dual-Agent Evidence Reasoning for Compact LLMs on Literary Long Narratives", "summary": "Researchers introduced ClueWeaver, a reward-guided dual-agent framework that improves question answering on long literary narratives using compact local language models. The framework, detailed in arXiv:2608.25531v1, uses a Finder agent to locate answer-critical passages and an Interpreter agent to derive answers with paragraph-ID citations, both optimized via reinforcement learning. Experiments show substantial gains over end-to-end local models, with code available on GitHub.", "body_md": "arXiv:2608.25531v1 Announce Type: new\nAbstract: Humanities and social science research requires close reading of long narrative materials such as novels, scripts, archives, and case reports, yet many users have limited access to costly proprietary long-context models. Compact, locally deployable language models are a practical alternative, but directly feeding them an entire long context remains costly, hard to inspect, and prone to missing sparse evidence. We present ClueWeaver, an evidence-aware dual-agent framework for long-narrative question answering with compact local models. A Finder identifies passages containing answer-critical clues through retrieval-guided segmentation, while an Interpreter derives the answer from the selected evidence, produces rationales with paragraph-ID citations, and applies an internal self-calibration pass for high-risk questions. Both agents are optimized with reward-guided reinforcement learning: Finder rewards emphasize evidence retention and faithful paragraph-ID references, and Interpreter rewards emphasize correctness, grounding, and concise explanations. This decomposition makes evidence selection and reasoning more inspectable than end-to-end prompting. Experiments across multiple long-context narrative question answering and claim verification settings show that ClueWeaver substantially improves local end-to-end language models while providing evidence coverage and paragraph-referenced reasoning traces. Code is available at https://github.com/Ameame1/ClueWeaver.", "url": "https://wpnews.pro/news/clueweaver-reward-guided-dual-agent-evidence-reasoning-for-compact-llms-on-long", "canonical_source": "https://www.machinebrief.com/news/clueweaver-reward-guided-dual-agent-evidence-reasoning-for-c-zn61", "published_at": "2026-08-27 04:00:00+00:00", "updated_at": "2026-08-27 05:18:58.198095+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "large-language-models", "ai-research"], "entities": ["ClueWeaver", "arXiv", "GitHub"], "alternates": {"html": "https://wpnews.pro/news/clueweaver-reward-guided-dual-agent-evidence-reasoning-for-compact-llms-on-long", "markdown": "https://wpnews.pro/news/clueweaver-reward-guided-dual-agent-evidence-reasoning-for-compact-llms-on-long.md", "text": "https://wpnews.pro/news/clueweaver-reward-guided-dual-agent-evidence-reasoning-for-compact-llms-on-long.txt", "jsonld": "https://wpnews.pro/news/clueweaver-reward-guided-dual-agent-evidence-reasoning-for-compact-llms-on-long.jsonld"}}