[2511.08579] Training Language Models to Explain Their Own Computations
Abstract:Can language models (LMs) learn to faithfully describe their internal computations? Are they better able to describe themselves than other models? We study the extent to which LMs' privileged access to their own internals can be leveraged to produce new techniques for explaining their behavior. Using existing interpretability techniques as a source of ground truth, we fine-tune LMs to generate natural language descriptions of (1) the information encoded by LM features, (2) the causal structure of LMs' internal activations, and (3) the influence of specific input tokens on LM outputs. When trained with only tens of thousands of example explanations, explainer models exhibit non-trivial generalization to new queries. This generalization appears partly attributable to explainer models' privileged access to their own internals: using a model to explain its own computations generally works better than using a *different* model to explain its computations (even if the explainer model is significantly more capable than the target). Our results suggest not only that LMs can learn to reliably explain their internal computations, but that such explanations offer a scalable complement to existing interpretability methods. Code and data at this https URL
Training Language Models to Explain Their Own Computations Belinda Z. Li 1 2 Zifan Carl Guo 2 Vincent Huang 1 Jacob Steinhardt 1 Jacob Andreas 1 2 Abstract factors that determine their decisions (Turpin et al., 2023; Chen et al., 2025; Barez et al., 2025). Can language models (LMs)…
saved by
related reading
- Training Language Models to Explain Their Own Computationsarxiv.org
- Training Language Models to Explain Their Own Computationsarxiv.org
- [2607.14111] Introspection Fine-Tuning (IFT): Training Small LLMs to Introspectarxiv.org
- Paper Summary: Interpreting Language Model Parameters - Goodfiregoodfire.com
- Language models can explain neurons in language modelsopenaipublic.blob.core.windows.net
- Natural Language Autoencoders Produce Unsupervised Explanations of LLM Activationstransformer-circuits.pub
- Transformer Circuits Threadtransformer-circuits.pub
- Self-CTRL: Self-Consistency Training with Reinforcement Learningarxiv.org
- On the Biology of a Large Language Modeltransformer-circuits.pub
- Tracing the thoughts of a large language model \ Anthropicanthropic.com
- Circuit Tracing: Revealing Computational Graphs in Language Modelstransformer-circuits.pub
- [2505.17120] Self-Interpretability: LLMs Can Describe Complex Internal Processes that Drive Their Decisionsarxiv.org