The Truthfulness Spectrum Hypothesis
Paper • 2602.20273 • Published • 2
Ignacio Roldan Fernandez PRO
igrofe
·
AI & ML interests
Aerospace engineer transitioning to technical AI safety research. Currently working on LLM honesty, deception, evaluations and context-dependent linear representations. Interests include AI control, scalable oversight and mechanistic interpretability.
Organizations
None yet
upvoted a paper 6 months ago