LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations
Paper
•
2410.02707
•
Published
•
49
•
5
Hadas Orgad
hadasor
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
13 days ago
CRISP: Persistent Concept Unlearning via Sparse Autoencoders
updated
a model
15 days ago
hadasor/protected_model_1
published
a model
15 days ago
hadasor/protected_model_1
Organizations
commented
2
papers
11 months ago