Chained Recursive Language Models for Multi-Iteration Reasoning
- ID: 98ab5582
- Original URL: https://arxiv.org/abs/2608.05124
- PDF: https://arxiv.org/pdf/2608.05124v1
- Author(s): Purbesh Mitra, Sennur Ulukus
- Date: 2026-08-05
- Category: cs.CL, cs.AI, cs.IT, cs.LG, eess.SP
- Source type: paper
- Tags: recursive-reasoning, context-management, artifact-workspace, long-context, multi-hop-reasoning
- Quality score: 4/5
- Fetched at: 2026-08-07T04:20:20+00:00
- Obsidian evidence: OpenCLI arXiv metadata backfill
中文导读
Chained Recursive Language Models 提出一种推理时架构：同一模型被重复调用为一串新的推理根，每个 root 都看到原问题和上下文，但不继承完整对话历史，而是接收简短摘要blackboard 和前序 root 写下的任务产物这是把长上下文推理拆成可检查可修正可续写的分段计算，适合抽取计数排序和 multi-hop 任务
为什么值得关注
This fills a high-score (4/5) AAIF content gap around recursive-reasoning, context-management, artifact-workspace, long-context, with the abstract giving enough grounded detail for follow-up reading and comparison.
English Summary
Long context reasoning in large language models (LLMs) is usually constrained by the fact that a single inference trajectory has to simultaneously explore the context, store intermediate state, verify evidence, and produce the final answer. This becomes particularly difficult in tasks that require extraction, counting, ordering, or multi-hop reasoning, where an early mistake can propagate until the final response. In this work, we propose Chained Recursive Language Models (Chained RLM), an inference-time architecture, in which the same underlying model is called repeatedly as a sequence of fresh reasoning roots. Each root receives the original problem and context, but does not inherit the full conversational history. Instead, it receives a compact plain-text summary, a plain-text blackboard, and some durable task-specific artifacts written by predecessor roots....
原文摘要 / Source Excerpt
Chained Recursive Language Models for Multi-Iteration Reasoning
- arXiv: https://arxiv.org/abs/2608.05124
- PDF: https://arxiv.org/pdf/2608.05124v1
- Authors: Purbesh Mitra, Sennur Ulukus
- Published: 2026-08-05
- Updated: 2026-08-05
- Categories: cs.CL, cs.AI, cs.IT, cs.LG, eess.SP
Abstract
Long context reasoning in large language models (LLMs) is usually constrained by the fact that a single inference trajectory has to simultaneously explore the context, store intermediate state, verify evidence, and produce the final answer. This becomes particularly difficult in tasks that require extraction, counting, ordering, or multi-hop reasoning, where an early mistake can propagate until the final response. In this work, we propose Chained Recursive Language Models (Chained RLM), an inference-time architecture, in which the same underlying model is called repeatedly as a sequence of fresh reasoning roots. Each root receives the original problem and context, but does not inherit the full conversational history. Instead, it receives a compact plain-text summary, a plain-text blackboard, and some durable task-specific artifacts written by predecessor roots. The motivation is to manage the context by chopping into partial tasks rather than one large inference response; in each staged computation, intermediate artifacts can be inspected, corrected, and extended by a later fresh inference by the same model. We describe the system model, handoff mechanism, artifact workspace, and evaluation protocol for this system. We study when fresh-context artifact continuation gives a measurable gain in accuracy over direct LLM answering even with recursive tool-calling.