LogitLens4LLMs: Extending Logit Lens Analysis to Modern Large Language Models
This work addresses the need for interpretability tools in AI research by enabling deeper investigations into the internal mechanisms of large-scale language models, though it is incremental as it builds on an existing technique.
This paper tackles the problem of applying Logit Lens analysis to modern large language models by introducing LogitLens4LLMs, a toolkit that extends the technique to state-of-the-art architectures like Qwen-2.5 and Llama-3.1, achieving full compatibility with the HuggingFace transformer library and low inference overhead.
This paper introduces LogitLens4LLMs, a toolkit that extends the Logit Lens technique to modern large language models. While Logit Lens has been a crucial method for understanding internal representations of language models, it was previously limited to earlier model architectures. Our work overcomes the limitations of existing implementations, enabling the technique to be applied to state-of-the-art architectures (such as Qwen-2.5 and Llama-3.1) while automating key analytical workflows. By developing component-specific hooks to capture both attention mechanisms and MLP outputs, our implementation achieves full compatibility with the HuggingFace transformer library while maintaining low inference overhead. The toolkit provides both interactive exploration and batch processing capabilities, supporting large-scale layer-wise analyses. Through open-sourcing our implementation, we aim to facilitate deeper investigations into the internal mechanisms of large-scale language models. The toolkit is openly available at https://github.com/zhenyu-02/LogitLens4LLMs.