Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
167 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Towards Measuring Representational Similarity of Large Language Models (2312.02730v1)

Published 5 Dec 2023 in cs.LG and cs.CL

Abstract: Understanding the similarity of the numerous released LLMs has many uses, e.g., simplifying model selection, detecting illegal model reuse, and advancing our understanding of what makes LLMs perform well. In this work, we measure the similarity of representations of a set of LLMs with 7B parameters. Our results suggest that some LLMs are substantially different from others. We identify challenges of using representational similarity measures that suggest the need of careful study of similarity scores to avoid false conclusions.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (48)
  1. Blackbox meets blackbox: Representational similarity & stability analysis of neural language models and brains. In Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 191–203, Florence, Italy. Association for Computational Linguistics.
  2. Pythia: A suite for analyzing large language models across training and scaling. In International Conference on Machine Learning, pages 2397–2430. PMLR.
  3. Bloom: A 176b-parameter open-access multilingual language model. ArXiv preprint, abs/2211.05100.
  4. Understanding the inner workings of language models through representation dissimilarity. ArXiv preprint, abs/2310.14993.
  5. Evaluating large language models trained on code. ArXiv preprint, abs/2107.03374.
  6. What does BERT look at? an analysis of BERT’s attention. In Linzen, T., Chrupała, G., Belinkov, Y., and Hupkes, D., editors, Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 276–286, Florence, Italy. Association for Computational Linguistics.
  7. Algorithms for learning kernels based on centered alignment. The Journal of Machine Learning Research, 13(1):795–828.
  8. On kernel-target alignment. In Dietterich, T., Becker, S., and Ghahramani, Z., editors, Advances in Neural Information Processing Systems, volume 14. MIT Press.
  9. BERT: Pre-training of deep bidirectional transformers for language understanding. In Burstein, J., Doran, C., and Solorio, T., editors, Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171–4186, Minneapolis, Minnesota. Association for Computational Linguistics.
  10. Grounding representation similarity through statistical testing. In Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J. W., editors, Advances in Neural Information Processing Systems, volume 34, pages 1556–1568. Curran Associates, Inc.
  11. A Mathematical Framework for Transformer Circuits. Transformer Circuits Thread.
  12. Ethayarajh, K. (2019). How contextual are contextualized word representations? Comparing the geometry of BERT, ELMo, and GPT-2 embeddings. In Inui, K., Jiang, J., Ng, V., and Wan, X., editors, Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 55–65, Hong Kong, China. Association for Computational Linguistics.
  13. OpenLLaMA: An Open Reproduction of LLaMA.
  14. Measuring Statistical Dependence with Hilbert-Schmidt Norms. In Jain, S., Simon, H. U., and Tomita, E., editors, Algorithmic Learning Theory, Lecture Notes in Computer Science, pages 63–77, Berlin, Heidelberg. Springer.
  15. Studying large language model generalization with influence functions. ArXiv preprint, abs/2308.03296.
  16. Finding neurons in a haystack: Case studies with sparse probing. ArXiv preprint, abs/2305.01610.
  17. Diachronic word embeddings reveal statistical laws of semantic change. In Erk, K. and Smith, N. A., editors, Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1489–1501, Berlin, Germany. Association for Computational Linguistics.
  18. Similarity of Neural Network Models: A Survey of Functional and Representational Measures. ArXiv preprint, abs/2305.06329.
  19. Large language models are zero-shot reasoners. In Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., and Oh, A., editors, Advances in Neural Information Processing Systems, volume 35, pages 22199–22213. Curran Associates, Inc.
  20. Similarity of neural network representations revisited. In Chaudhuri, K. and Salakhutdinov, R., editors, Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceedings of Machine Learning Research, pages 3519–3529. PMLR.
  21. Representational similarity analysis - connecting the branches of systems neuroscience. Frontiers in Systems Neuroscience, 2.
  22. Linguistic knowledge and transferability of contextual representations. In Burstein, J., Doran, C., and Solorio, T., editors, Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 1073–1094, Minneapolis, Minnesota. Association for Computational Linguistics.
  23. BERTs of a feather do not generalize together: Large variability in generalization across models with similar test set performance. In Alishahi, A., Belinkov, Y., Chrupała, G., Hupkes, D., Pinter, Y., and Sajjad, H., editors, Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP, pages 217–227, Online. Association for Computational Linguistics.
  24. MosaicML (2023). Introducing MPT-7B: A New Standard for Open-Source, Commercially Usable LLMs.
  25. Relative representations enable zero-shot latent space communication. In The Eleventh International Conference on Learning Representations.
  26. The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora with Web Data, and Web Data Only. ArXiv preprint, abs/2306.01116.
  27. Deep contextualized word representations. In Walker, M., Ji, H., and Stent, A., editors, Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 2227–2237, New Orleans, Louisiana. Association for Computational Linguistics.
  28. Code Llama: Open Foundation Models for Code. ArXiv preprint, abs/2308.12950.
  29. Winogrande: An adversarial winograd schema challenge at scale. Proceedings of the AAAI Conference on Artificial Intelligence, 34(05):8732–8740.
  30. The Effects of Randomness on the Stability of Node Embeddings. In Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Communications in Computer and Information Science, pages 197–215. Springer International Publishing.
  31. ModelDiff: A framework for comparing learning algorithms. In Krause, A., Brunskill, E., Cho, K., Engelhardt, B., Sabato, S., and Scarlett, J., editors, Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 30646–30688. PMLR.
  32. Using distance on the riemannian manifold to compare representations in brain and in models. NeuroImage, 239:118271.
  33. Beyond the imitation game: Quantifying and extrapolating the capabilities of language models. Transactions on Machine Learning Research.
  34. StabilityAI (2023). StableLM Alpha.
  35. Galactica: A Large Language Model for Science. ArXiv preprint, abs/2211.09085.
  36. All bark and no bite: Rogue dimensions in transformer language models obscure representational quality. In Moens, M.-F., Huang, X., Specia, L., and Yih, S. W.-t., editors, Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 4527–4546, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  37. Together.ai (2023). RedPajama 7B now available, instruct model outperforms all open 7B models on HELM benchmarks.
  38. LLaMA: Open and Efficient Foundation Language Models. ArXiv preprint, abs/2302.13971.
  39. GPT-J-6B: A 6 Billion Parameter Autoregressive Language Model. https://github.com/kingoflolz/mesh-transformer-jax.
  40. Towards understanding the instability of network embedding. IEEE Transactions on Knowledge and Data Engineering, 34(2):927–941.
  41. Understanding Weight Similarity of Neural Networks via Chain Normalization Rule and Hypothesis-Training-Testing. ArXiv preprint, abs/2208.04369.
  42. Chain-of-thought prompting elicits reasoning in large language models. In Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., and Oh, A., editors, Advances in Neural Information Processing Systems, volume 35, pages 24824–24837. Curran Associates, Inc.
  43. Similarity analysis of contextual word representation models. In Jurafsky, D., Chai, J., Schluter, N., and Tetreault, J., editors, Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4638–4655, Online. Association for Computational Linguistics.
  44. Harnessing the Power of LLMs in Practice: A Survey on ChatGPT and Beyond. ArXiv preprint, abs/2304.13712.
  45. On the dimensionality of word embedding. In Bengio, S., Wallach, H., Larochelle, H., Grauman, K., Cesa-Bianchi, N., and Garnett, R., editors, Advances in Neural Information Processing Systems, volume 31. Curran Associates, Inc.
  46. In-Context Learning in Large Language Models: A Neuroscience-inspired Analysis of Representations.
  47. OPT: Open Pre-trained Transformer Language Models. ArXiv preprint, abs/2205.01068.
  48. A Survey of Large Language Models. ArXiv preprint, abs/2303.18223.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
Github Logo Streamline Icon: https://streamlinehq.com

GitHub

  1. GitHub - mklabunde/llm_repsim (4 stars)