4 papers
Evil Spectra: How Optimisers can Amplify or Suppress Emergent Misalignment
Jason R. Brown, Patrick Leask, Lev McKinney
Emergent misalignment (EM) is a recently discovered phenomenon in LLMs where fine-tuning on a narrow misaligned task, such as writing insecure code, leads to broadly misaligned beh…
Concept Influence: Leveraging Interpretability to Improve Performance and Efficiency in Training Data Attribution
Matthew Kowal, Goncalo Paulo, Louis Jaburi +6
As large language models are increasingly trained and fine-tuned, practitioners need methods to identify which training data drive specific behaviors, particularly unintended ones.…
Gauss-Newton Unlearning for the LLM Era
Lev McKinney, Anvith Thudi, Juhan Bae +4
Standard large language model training can create models that produce outputs their trainer deems unacceptable in deployment. The probability of these outputs can be reduced using…
Eliciting Latent Predictions from Transformers with the Tuned Lens
Nora Belrose, Igor Ostrovsky, Lev McKinney +5
We analyze transformers from the perspective of iterative inference, seeking to understand how model predictions are refined layer by layer. To do so, we train an affine probe for…