RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations
Jing Huang, Zhengxuan Wu, Christopher Potts, Mor Geva, Atticus Geiger
Abstract
Individual neurons participate in the representation of multiple high-level concepts. To what extent can different interpretability methods successfully disentangle these roles? To help address this question, we introduce RAVEL (Resolving Attribute-Value Entanglements in Language Models), a dataset that enables tightly controlled, quantitative comparisons between a variety of existing interpretability methods. We use the resulting conceptual framework to define the new method of Multi-task Distributed Alignment Search (MDAS), which allows us to find distributed representations satisfying multiple causal criteria. With Llama2-7B as the target language model, MDAS achieves state-of-the-art results on RAVEL, demonstrating the importance of going beyond neuron-level analyses to identify features distributed across activations. We release our benchmark at https://github.com/explanare/ravel.- Anthology ID:
- 2024.acl-long.470
- Volume:
- Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Month:
- August
- Year:
- 2024
- Address:
- Bangkok, Thailand
- Editors:
- Lun-Wei Ku, Andre Martins, Vivek Srikumar
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 8669–8687
- Language:
- URL:
- https://aclanthology.org/2024.acl-long.470
- DOI:
- Cite (ACL):
- Jing Huang, Zhengxuan Wu, Christopher Potts, Mor Geva, and Atticus Geiger. 2024. RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 8669–8687, Bangkok, Thailand. Association for Computational Linguistics.
- Cite (Informal):
- RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations (Huang et al., ACL 2024)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-4/2024.acl-long.470.pdf