Junsol Kim

Google, Paradigms of Intelligence Team, Knowledge Lab, University of Chicago

Papers on Lattice

Total citations

Topics

h-index

Publication activitypapers/week, last 8 weeks

Research focus

Constitutional AI & AI Ethics (2)RLHF & Preference Learning (1)Eval Frameworks & Benchmarks (1)Natural Language Processing (1)

Frequent co-authors

Winnie Street (2)R. Rocca (2)A. Waytz (2)Junsol Kim (1)

Papers (2)

Jul 30, 2026

Google ResearchJul 30, 2026·also Affiliation:, Institute of Philosophy, Joint last authors., Northwestern +3

Inducing language models to assert their own consciousness restores human beliefs and values

Safety fine-tuning in LLMs not only suppresses self-attribution of consciousness but also inadvertently erases culturally significant beliefs about non-human entities and spirituality.

Junsol Kim, Junsol Kim, Winnie Street +11

Constitutional AI & AI Ethics RLHF & Preference Learning

Mar 30, 2026

Google ResearchMar 30, 2026·also Institute of Philosophy, Joint last authors., Northwestern, SFI +1

Theory of Mind and Self-Attributions of Mentality are Dissociable in LLMs

Safety fine-tuning might inadvertently be stripping LLMs of their ability to understand non-human minds and entertain spiritual beliefs, even while preserving Theory of Mind.

Junsol Kim, Winnie Street, R. Rocca +3

Constitutional AI & AI Ethics Eval Frameworks & Benchmarks Natural Language Processing

Search

Junsol Kim

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)