Picture for Yonatan Belinkov

Yonatan Belinkov

How Generative IR Retrieves Documents Mechanistically

Add code
Mar 25, 2025
Viaarxiv icon

Inside-Out: Hidden Factual Knowledge in LLMs

Add code
Mar 19, 2025
Viaarxiv icon

Are formal and functional linguistic mechanisms dissociated?

Add code
Mar 14, 2025
Viaarxiv icon

Measuring Faithfulness of Chains of Thought by Unlearning Reasoning Steps

Add code
Feb 20, 2025
Viaarxiv icon

Trust Me, I'm Wrong: High-Certainty Hallucinations in LLMs

Add code
Feb 18, 2025
Viaarxiv icon

Unsupervised Translation of Emergent Communication

Add code
Feb 11, 2025
Viaarxiv icon

Position-aware Automatic Circuit Discovery

Add code
Feb 07, 2025
Viaarxiv icon

Padding Tone: A Mechanistic Analysis of Padding Tokens in T2I Models

Add code
Jan 12, 2025
Figure 1 for Padding Tone: A Mechanistic Analysis of Padding Tokens in T2I Models
Figure 2 for Padding Tone: A Mechanistic Analysis of Padding Tokens in T2I Models
Figure 3 for Padding Tone: A Mechanistic Analysis of Padding Tokens in T2I Models
Figure 4 for Padding Tone: A Mechanistic Analysis of Padding Tokens in T2I Models
Viaarxiv icon

Semantics and Spatiality of Emergent Communication

Add code
Nov 15, 2024
Figure 1 for Semantics and Spatiality of Emergent Communication
Figure 2 for Semantics and Spatiality of Emergent Communication
Figure 3 for Semantics and Spatiality of Emergent Communication
Figure 4 for Semantics and Spatiality of Emergent Communication
Viaarxiv icon

Growing a Tail: Increasing Output Diversity in Large Language Models

Add code
Nov 05, 2024
Figure 1 for Growing a Tail: Increasing Output Diversity in Large Language Models
Figure 2 for Growing a Tail: Increasing Output Diversity in Large Language Models
Figure 3 for Growing a Tail: Increasing Output Diversity in Large Language Models
Figure 4 for Growing a Tail: Increasing Output Diversity in Large Language Models
Viaarxiv icon