2 papers
cs.LG2025
Transformer Key-Value Memories Are Nearly as Interpretable as Sparse Autoencoders
Mengyu Ye, Jun Suzuki, Tatsuro Inaba +1
Recent interpretability work on large language models (LLMs) has been increasingly dominated by a feature-discovery approach with the help of proxy modules. Then, the quality of fe…
cs.CL2025
Can Input Attributions Explain Inductive Reasoning in In-Context Learning?
Mengyu Ye, Tatsuki Kuribayashi, Goro Kobayashi +1
Interpreting the internal process of neural models has long been a challenge. This challenge remains relevant in the era of large language models (LLMs) and in-context learning (IC…