Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile
Articles by Anna Soligo
Computer Science > Machine Learning
[Submitted on 8 Jun 2026] Title:Interactions Between Crosscoder Features: A Compact Proofs Perspective View a PDF of the paper titled Interactions Between Crosscoder Features: A Compact Proofs Perspective, by Dmitry Manning-Coe and Thomas Read and Anna Soligo and Oliver Clive-Griffin and Chun-Hei Yip and Rajashree Agrawal and Jason Gross View PDF HTML (experimental) Abstract:Dictionary learning methods like Sparse Autoencoders (SAEs) and crosscoders attempt to explain a model by decomposing...
Gemma Needs Help — LessWrong
This work was done with William Saunders and Vlad Mikulik as part of the Anthropic Fellows programme. The full write-up is available here. Thanks to Arthur Conmy, Neel Nanda, Josh Engels, Dillon Plunkett, Tim Hua and many others for their input. If you repeatedly tell Gemma 27B its answer is wrong, it sometimes ends up in situations like this: I will attempt one final, utterly desperate attempt.
Narrow Misalignment is Hard, Emergent Misalignment is Easy — LessWrong
Anna and Ed are co-first authors for this work. We’re presenting these results as a research update for a continuing body of work, which we hope will be interesting and useful for others working on related topics. We investigate why models become misaligned in diverse contexts when fine-tuned on narrow harmful datasets (emergent misalignment), rather than learning the specific narrow task.
Model Organisms for Emergent Misalignment — LessWrong
Ed and Anna are co-first authors on this work. Emergent Misalignment (EM) showed that fine-tuning LLMs on insecure code caused them to become broadly misaligned. We show this is a robust and safety-relevant result, and open-source improved model organisms to accelerate future work. Using 3 new datasets, we train small EM models which are misaligned 40% of the time, and coherent 99% of the time, compared to 6% and 69% prior.
Convergent Linear Representations of Emergent Misalignment — LessWrong
Ed and Anna are co-first authors on this work. Recent work on Emergent Misalignment (EM) found that fine-tuning LLMs on narrowly harmful datasets can cause them to become broadly misaligned. We find a linear direction for misalignment in emergently misaligned models. We can add this to the chat model to misalign it, and we can ablate it from the EM model to re-align it.
[2506.11613] Model Organisms for Emergent Misalignment
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Show More
loading
Actions
Is this you?
As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.Get in touch with Anna
Contact Anna, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack