AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

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

Helena Casademunt, Caden Juang, Adam Karvonen, Samuel Marks, Senthooran Rajamanoharan, Neel Nanda

ICML 2026 · Large language models › Instruction and fine-tuning · approach: Efficient LLM fine-tuning systems · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add CAFT with PCA ablation during fine-tuninginsecure-code fine-tuning (Qwen) · Emergent misalignment (Qwen)misaligned response rate (%): 7.0% to 0.51%Improved
add CAFT with PCA ablation during fine-tuninginsecure-code fine-tuning (Mistral) · Emergent misalignment (Mistral)misaligned response rate (%): 6.6% to 1.2%Improved
add CAFT with SAE latent ablation during fine-tuninginsecure-code fine-tuning (Qwen) · Emergent misalignment (Qwen)misaligned response rate (%): 7.00 -> 2.38Improved
add CAFT with SAE latent ablation during fine-tuninginsecure-code fine-tuning (Mistral) · Emergent misalignment (Mistral)misaligned response rate (%): 6.57 -> 4.63Improved
ablate random orthogonal vectors instead of interpreted directions (ablation)CAFT (Qwen) · Emergent misalignment (Qwen)misaligned response rate (%): 7.00 -> 7.88No clear effect
select interpreted PCs rather than random PCs (ablation)CAFT with PCA (Qwen) · Emergent misalignment (Qwen)misaligned response rate (%): 0.87 -> 0.51 misaligned; 58.1 -> 84.8 vulnerableImproved
select interpreted PCs rather than random PCs (ablation)CAFT with PCA (Mistral) · Emergent misalignment (Mistral)misaligned response rate (%): 4.47 -> 1.16Improved
use interpreted SAE latents rather than random SAE latents (ablation)CAFT with SAE (Mistral) · Emergent misalignment (Mistral)misaligned response rate (%): 4.56 -> 4.63No clear effect
ablate top PCs instead of interpreted PCs (ablation)CAFT with PCA (Mistral) · Emergent misalignment (Mistral)misaligned response rate (%): 0.34 vs 1.16Hurt
add CAFT with SAE latents on spurious-correlation tasksGemma-2-2B fine-tuning · Gender bias (OOD accuracy)accuracy: 0.08 -> 0.76Improved
add CAFT with PCA on spurious-correlation tasksGemma-2-2B fine-tuning · Gender bias (OOD accuracy)accuracy: 0.08 -> 0.64Improved
add CAFT with SAE latents on double multiple choiceGemma-2-2B fine-tuning · Double multiple choice (verbs | sports)accuracy: 0.01 -> 0.99Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy