[2604.16812] Introspection Adapters: Training LLMs to Report Their Learned Behaviors

Abstract:When model developers or users fine-tune an LLM, this can induce behaviors that are unexpected, deliberately harmful, or hard to detect. It would be far easier to audit LLMs if they could simply describe their behaviors in natural language. Here, we study a scalable approach to rapidly identify learned behaviors of many LLMs derived from a shared base LLM. Given a model $M$, our method works by finetuning models $M_i$ from $M$ with implanted behaviors $b_i$; the $(M_i, b_i)$ pairs serve as labeled training data. We then train an introspection adapter (IA): a single LoRA adapter jointly trained across the finetunes $M_i$ to cause them to verbalize their implanted behaviors. We find that this IA induces self-description of learned behaviors even in finetunes of $M$ that were trained in very different ways from the $M_i$. For example, IAs generalize to AuditBench, achieving state-of-the-art at identifying explicitly hidden concerning behaviors. IAs can also be used to detect encrypted finetuning API attacks. They scale favorably with model size and training data diversity. Overall, our results suggest that IAs are a scalable, effective, and practically useful approach to auditing fine-tuned LLMs.

View PDF HTML (experimental) Abstract:When model developers or users fine-tune an LLM, this can induce behaviors that are unexpected, deliberately harmful, or hard to detect. It would be far easier to audit LLMs if they could simply describe their behaviors in natural language. Here, we study a scalable approach to rapidly identify learned behaviors of many LLMs derived from a shared base LLM. Given a model $M$, our method works by finetuning models $M_i$ from $M$ with implanted behaviors $b_i$; the $(M_i, b_i)$ pairs serve as labeled training data. We then train an introspection adapter…

saved by

• Yixiong Hao
• Arya Pasumarthi
• Ishaan Panigrahi
• Emanuel Ruzak

related reading

• [2604.16812] Introspection Adapters: Training LLMs to Report Their Learned Behaviorsarxiv.org
• Introspection Adapters: Training LLMs to Report Their Learned Behaviorsarxiv.org
• Spilling the Beans: Teaching LLMs to Self-Report Their Hidden Objectivesarxiv.org
• Modifying LLM Beliefs with Synthetic Document Finetuningalignment.anthropic.com
• Probe-Based Data Attribution: Surfacing and Mitigating Undesirable Behaviors in LLM Post-Training - Goodfiregoodfire.com
• [2503.10965] Auditing language models for hidden objectivesarxiv.org
• [2607.14111] Introspection Fine-Tuning (IFT): Training Small LLMs to Introspectarxiv.org
• Auditing language models for hidden objectives — LessWronglesswrong.com
• [2511.08579] Training Language Models to Explain Their Own Computationsarxiv.org
• Emergent introspective awareness in large language models \ Anthropicanthropic.com
• Can LLMs Introspect? A Reality Checkarxiv.org
• Training Language Models to Explain Their Own Computationsarxiv.org