Scholar
Mario Giulianelli
Google Scholar ID: ABZghWYAAAAJ
Associate Professor, UCL
Citations & Impact
Citations
3,188
H-index
16
i10-index
20
Publications
20
Co-authors
23
list available
Publications
Resume (English only)
Academic Achievements
- Co-authored a 2018 paper introducing one of the first mechanistic interpretability methods for language models.
- Paper on exact and fast approximate algorithms for converting token-level to character-level language models accepted as ICML Spotlight (top 2.6%).
- Paper on inductive biases in human and neural language models selected for ACL panel discussion (top 0.8%).
- Four papers accepted at ACL and two at ICML.
- GenBench taxonomy paper published in Nature Machine Intelligence.
- Awarded a Cohere Labs Catalyst Grant for SPAR experiments on measuring goal-directedness in LLM agents.
- Awarded a Cosmos Grant to study emergent self-interest in LLM agents using non-cooperative dialogue games.
- Serving as a SPAR mentor.
- Led an international joint testing exercise on capability elicitation, LLM judges, and agentic testing methodology.
