4 papers · 1 filter
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…
TopK Language Models
Ryosuke Takahashi, Tatsuro Inaba, Kentaro Inui +1
Sparse autoencoders (SAEs) have become an important tool for analyzing and interpreting the activation space of transformer-based language models (LMs). However, SAEs suffer severa…
How a Bilingual LM Becomes Bilingual: Tracing Internal Representations with Sparse Autoencoders
Tatsuro Inaba, Go Kamoda, Kentaro Inui +5
This study explores how bilingual language models develop complex internal representations. We employ sparse autoencoders to analyze internal representations of bilingual language…
Weight-based Analysis of Detokenization in Language Models: Understanding the First Stage of Inference Without Inference
Go Kamoda, Benjamin Heinzerling, Tatsuro Inaba +3
According to the stages-of-inference hypothesis, early layers of language models map their subword-tokenized input, which does not necessarily correspond to a linguistically meanin…