Poster · Sydney
Uncovering Hidden Propensities in Language Models via Limited-Parameter Finetuning
- When
- Tue 8 Dec 2026, 10:00–13:00 AEDT
- Where
- Hall 1-4, ICC Sydney
Exhibition Centre, Level 1 - Session
- Sydney Poster Session 1
- Authors
- Boyd Kane, Jo Jiao, Bryce Woodworth, Elizabeth Donoway, Alex Cloud, Alexander Turner
- Organisations
- MATS Research, University of Chicago, University of California, Berkeley, Anthropic, Google DeepMind
- QRS classification
- Language models