An Investigation of Neuron Activation as a Unified Lens to Explain Chain-of-Thought Eliciting Arithmetic Reasoning of LLMs

Documenti analoghi

• Data-driven Circuit Discovery for Interpretability of Language Models
  di: Rai, Daking, et al.
  Pubblicazione: (2026)
• Mechanistic Understanding of Language Models in Syntactic Code Completion
  di: Miller, Samuel, et al.
  Pubblicazione: (2025)
• A Practical Review of Mechanistic Interpretability for Transformer-Based Language Models
  di: Rai, Daking, et al.
  Pubblicazione: (2024)
• Failure by Interference: Language Models Make Balanced Parentheses Errors When Faulty Mechanisms Overshadow Sound Ones
  di: Rai, Daking, et al.
  Pubblicazione: (2025)
• All for One: LLMs Solve Mental Math at the Last Token With Information Transferred From Other Tokens
  di: Mamidanna, Siddarth, et al.
  Pubblicazione: (2025)