Bootstrapping RLM-Specific Reasoners
Fine-tune smaller open-source language models on successful high-efficiency recursive language model trajectories to improve context management.
Source paper
Generated by AI. The paper’s authors did not write these.
Fine-tune smaller open-source language models on successful high-efficiency recursive language model trajectories to improve context management.
Extend the recursive language model environment with a runtime stack and recursion logic to support deep multi-level sub-calls.
Refactor the recursive language model runtime with asynchronous primitives so sub-language models can execute chunk processing jobs in parallel.
Enforce strict REPL output schemas via grammar-based constrained decoding to eliminate brittle text-based tag parsing.
This paper introduces Recursive Language Models (RLMs), an inference-time scaling strategy that allows Large Language Models to process arbitrarily long inputs by treating the context as an external environment within a Python REPL. Instead of ingesting the full context at once, the model writes code to programmatically access, decompose, and recursively sub-query specific chunks of the data, enabling effective processing of inputs up to 10M+ tokens with competitive costs compared to base models.
The core methodology involves wrapping an LLM in a Read-Eval-Print Loop (REPL) environment where the long context is stored as a variable. The model is prompted to interact with this variable via Python code and a special `llm_query` function, allowing it to perform recursive decomposition (chunking) and sub-calls (processing chunks) to synthesize an answer without ever fitting the whole context into its immediate attention window.