Fabian Degen
Anthropic Research Fellow (contract)
Alignment research
I'm an Anthropic Research Fellow (contract) working on alignment research. Before that I did an MSc in Advanced Computer Science at the University of Oxford, where my thesis was primarily supervised by Jialin Yu and co-supervised by Philip Torr and Yarin Gal. In my thesis I built a benchmark for counterfactual prediction with LLMs and process verifiers that catch and correct the models' errors. I also measured which errors the verifiers cannot catch, and how wrong an attacker who knows the verifiers can make an answer without being caught.
1Publications
Plan2Map is a 208-case benchmark: given only a UK planning document, a system must reconstruct the machine-readable geospatial boundary it describes. I built GeoPlanAgent, the tool-augmented agent system introduced in the paper — it reads the document, locates the site, registers the map against a national basemap, segments the drawn boundary with a fine-tuned SAM 3, and projects it to geographic coordinates, reaching 0.736 mean IoU where direct VLM baselines reach 0.108.
2Open source
Supported by a Manifund grant, I helped build the TransformerLens 3.0 release, which grew the number of supported models from ~200 to ~9,000 and brought model outputs in line with the Hugging Face reference implementations; my main contributions were the backwards-compatibility layer and weight-level Q/K/V splitting.
Languages: German, English, some Spanish.