# Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning

arXiv: https://arxiv.org/abs/2507.16795 · HTML: https://mlsanity.com/paper/2507.16795/

## Key facts

- Cited by 39 papers in the map (data through 14 Oct 2026).
- Builds on 60 of its 90 references that are in the map.
- First posted on arXiv 22 Jul 2025.

## Builds on (most cited first, 30 of 60)

- [The Power of Scale for Parameter-Efficient Prompt Tuning](https://mlsanity.com/paper/2104.08691/) (2021)
- [Distributed Representations of Words and Phrases and their Compositionality](https://mlsanity.com/paper/1310.4546/) (2013)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings](https://mlsanity.com/paper/1607.06520/) (2016)
- [The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale](https://mlsanity.com/paper/2406.17557/) (2024)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [Discovering Language Model Behaviors with Model-Written Evaluations](https://mlsanity.com/paper/2212.09251/) (2022)
- [Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution](https://mlsanity.com/paper/2202.10054/) (2022)
- [The Linear Representation Hypothesis and the Geometry of Large Language Models](https://mlsanity.com/paper/2311.03658/) (2023)
- [The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets](https://mlsanity.com/paper/2310.06824/) (2023)
- [Editing Factual Knowledge in Language Models](https://mlsanity.com/paper/2104.08164/) (2021)
- [Just Train Twice: Improving Group Robustness without Training Group Information](https://mlsanity.com/paper/2107.09044/) (2021)
- [Fast Model Editing at Scale](https://mlsanity.com/paper/2110.11309/) (2021)
- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection](https://mlsanity.com/paper/2004.07667/) (2020)
- [Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models](https://mlsanity.com/paper/2403.19647/) (2024)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [Confounding variables can degrade generalization performance of radiological deep learning models](https://mlsanity.com/paper/1807.00431/) (2018)
- [Bias in Bios: A Case Study of Semantic Representation Bias in a High-Stakes Setting](https://mlsanity.com/paper/1901.09451/) (2019)
- [Finding Neurons in a Haystack: Case Studies with Sparse Probing](https://mlsanity.com/paper/2305.01610/) (2023)
- [Studying Large Language Model Generalization with Influence Functions](https://mlsanity.com/paper/2308.03296/) (2023)
- [Emergent Linear Representations in World Models of Self-Supervised Sequence Models](https://mlsanity.com/paper/2309.00941/) (2023)
- [Function Vectors in Large Language Models](https://mlsanity.com/paper/2310.15213/) (2023)
- [LEACE: Perfect linear concept erasure in closed form](https://mlsanity.com/paper/2306.03819/) (2023)
- [Learning Not to Learn: Training Deep Neural Networks with Biased Data](https://mlsanity.com/paper/1812.10352/) (2018)
- [ToxicChat: Unveiling Hidden Challenges of Toxicity Detection in Real-World User-AI Conversation](https://mlsanity.com/paper/2310.17389/) (2023)

## Built on by (most cited first, 30 of 39)

- [Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time](https://mlsanity.com/paper/2510.04340/) (2025)
- [Narrow Finetuning Leaves Clearly Readable Traces in Activation Differences](https://mlsanity.com/paper/2510.13900/) (2025)
- [Qwen-Scope: Turning Sparse Features into Development Tools for Large Language Models](https://mlsanity.com/paper/2605.11887/) (2026)
- [Steering Language Models with Weight Arithmetic](https://mlsanity.com/paper/2511.05408/) (2025)
- [Model Spec Midtraining: Improving How Alignment Training Generalizes](https://mlsanity.com/paper/2605.02087/) (2026)
- [Spilling the Beans: Teaching LLMs to Self-Report Their Hidden Objectives](https://mlsanity.com/paper/2511.06626/) (2025)
- [Hallucination reduction with CASAL: Contrastive Activation Steering For Amortized Learning](https://mlsanity.com/paper/2510.02324/) (2025)
- [Narrow Fine-Tuning Erodes Safety Alignment in Vision-Language Agents](https://mlsanity.com/paper/2602.16931/) (2026)
- [Delta-Crosscoder: Robust Crosscoder Model Diffing in Narrow Fine-Tuning Regimes](https://mlsanity.com/paper/2603.04426/) (2026)
- [Detecting Adversarial Fine-tuning with Auditing Agents](https://mlsanity.com/paper/2510.16255/) (2025)
- [BLOCK-EM: Preventing Emergent Misalignment via Latent Blocking](https://mlsanity.com/paper/2602.00767/) (2026)
- [One Bias After Another: Mechanistic Reward Shaping and Persistent Biases in Language Reward Models](https://mlsanity.com/paper/2603.03291/) (2026)
- [Reinforcement Learning Can Amplify Emergent Misalignment from Harmless Rewards](https://mlsanity.com/paper/2605.31328/) (2026)
- [Inoculation Adapters: Improved Selective Generalization of Capabilities with Fewer Surprising Backdoors](https://mlsanity.com/paper/2606.30252/) (2026)
- [Emergent Misalignment Recruits a Pre-existing Persona Subspace](https://mlsanity.com/paper/2607.21356/) (2026)
- [The Unintended Trade-off of AI Alignment:Balancing Hallucination Mitigation and Safety in LLMs](https://mlsanity.com/paper/2510.07775/) (2025)
- [Can SAEs reveal and mitigate racial biases of LLMs in healthcare?](https://mlsanity.com/paper/2511.00177/) (2025)
- [Convergent World Representations and Divergent Tasks](https://mlsanity.com/paper/2602.00533/) (2026)
- [Overtrained, Not Misaligned](https://mlsanity.com/paper/2605.12199/) (2026)
- [Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal](https://mlsanity.com/paper/2606.12360/) (2026)
- [Objective Matters: Fine-Tuning Objectives Shape Safety, Robustness, and Persona Drift](https://mlsanity.com/paper/2601.12639/) (2026)
- [Not Just the Destination, But the Journey: Reasoning Traces Causally Shape Generalization Behaviors](https://mlsanity.com/paper/2603.12397/) (2026)
- [Actionable Activation Directions for Detecting and Mitigating Emergent Misalignment Across Language Model Families](https://mlsanity.com/paper/2606.20225/) (2026)
- [Out-of-Distribution Generalization of Risk Aversion in Language Models](https://mlsanity.com/paper/2607.02755/) (2026)
- [From Weights to Words: Expressing and Editing Preference Model Inferences in Natural Language](https://mlsanity.com/paper/2607.16232/) (2026)
- [Inference-Time Consensus for Mitigating Hidden Behaviors from LLM Fine-Tuning](https://mlsanity.com/paper/2607.23394/) (2026)
- [Data Attribution of Emergent Misalignment with Persona Features](https://mlsanity.com/paper/2608.11025/) (2026)
- [The Unequal Influence of Bad Advice: Using Training Data Attribution to Modulate Emergent Misalignment](https://mlsanity.com/paper/2609.37914/) (2026)
- [MoRFI: Monotonic Sparse Autoencoder Feature Identification](https://mlsanity.com/paper/2604.26866/) (2026)
- [Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning](https://mlsanity.com/paper/2606.07631/) (2026)

## Abstract (arXiv)

Fine-tuning large language models (LLMs) can lead to unintended out-of-distribution generalization. Standard approaches to this problem rely on modifying training data, for example by adding data that better specify the intended generalization. However, this is not always practical. We introduce Concept Ablation Fine-Tuning (CAFT), a technique that leverages interpretability tools to control how LLMs generalize from fine-tuning, without needing to modify the training data or otherwise use data from the target distribution. Given a set of directions in an LLM's latent space corresponding to undesired concepts, CAFT works by ablating these concepts with linear projections during fine-tuning, steering the model away from unintended generalizations. We successfully apply CAFT to three fine-tuning tasks, including emergent misalignment, a phenomenon where LLMs fine-tuned on a narrow task generalize to give egregiously misaligned responses to general questions. Without any changes to the fine-tuning data, CAFT reduces misaligned responses by 10x without degrading performance on the training distribution. Overall, CAFT represents a novel approach for steering LLM generalization without modifying training data.