Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
126 tokens/sec
GPT-4o
47 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Experimental Contexts Can Facilitate Robust Semantic Property Inference in Language Models, but Inconsistently (2401.06640v2)

Published 12 Jan 2024 in cs.CL and cs.AI

Abstract: Recent zero-shot evaluations have highlighted important limitations in the abilities of LLMs (LMs) to perform meaning extraction. However, it is now well known that LMs can demonstrate radical improvements in the presence of experimental contexts such as in-context examples and instructions. How well does this translate to previously studied meaning-sensitive tasks? We present a case-study on the extent to which experimental contexts can improve LMs' robustness in performing property inheritance -- predicting semantic properties of novel concepts, a task that they have been previously shown to fail on. Upon carefully controlling the nature of the in-context examples and the instructions, our work reveals that they can indeed lead to non-trivial property inheritance behavior in LMs. However, this ability is inconsistent: with a minimal reformulation of the task, some LMs were found to pick up on shallow, non-semantic heuristics from their inputs, suggesting that the computational principles of semantic property inference are yet to be mastered by LMs.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (31)
  1. Back to square one: Artifact detection, training and commonsense disentanglement in the Winograd schema. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10486–10500, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  2. Allyson Ettinger. 2020. What BERT is not: Lessons from a new suite of psycholinguistic diagnostics for language models. Transactions of the Association for Computational Linguistics, 8:34–48.
  3. Neural language models as psycholinguistic subjects: Representations of syntactic state. In 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 32–42, Minneapolis, Minnesota. Association for Computational Linguistics.
  4. Colorless green recurrent networks dream hierarchically. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 1195–1205, New Orleans, Louisiana. Association for Computational Linguistics.
  5. How does GPT-2 compute greater-than?: Interpreting mathematical abilities in a pre-trained language model. In Thirty-seventh Conference on Neural Information Processing Systems.
  6. A systematic assessment of syntactic generalization in neural language models. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1725–1744, Online. Association for Computational Linguistics.
  7. Mistral 7b. arXiv preprint arXiv:2310.06825.
  8. Najoung Kim and Sebastian Schuster. 2023. Entity tracking in language models. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3835–3855, Toronto, Canada. Association for Computational Linguistics.
  9. Andrew Kyle Lampinen. 2022. Can language models handle recursively nested grammatical structures? a case study on comparing models and humans. arXiv preprint arXiv:2210.15303.
  10. Assessing the ability of LSTMs to learn syntax-sensitive dependencies. Transactions of the Association for Computational Linguistics, 4:521–535.
  11. Does syntax need to grow on trees? sources of hierarchical inductive bias in sequence-to-sequence networks. Transactions of the Association for Computational Linguistics, 8:125–140.
  12. Right for the wrong reasons: Diagnosing syntactic heuristics in natural language inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3428–3448, Florence, Italy. Association for Computational Linguistics.
  13. Kanishka Misra. 2022. minicons: Enabling flexible behavioral and representational analyses of transformer language models. arXiv preprint arXiv:2203.13112.
  14. COMPS: Conceptual minimal pair sentences for testing robust property knowledge and its inheritance in pre-trained language models. In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, pages 2928–2949, Dubrovnik, Croatia. Association for Computational Linguistics.
  15. Coloring the blank slate: Pre-training imparts a hierarchical inductive bias to sequence-to-sequence models. In Findings of the Association for Computational Linguistics: ACL 2022, pages 1352–1368, Dublin, Ireland. Association for Computational Linguistics.
  16. In-context learning generalizes, but not always robustly: The case of syntax. arXiv preprint arXiv:2311.07811.
  17. Gregory L Murphy. 2002. The Big Book of Concepts. MIT press.
  18. Lalchand Pandia and Allyson Ettinger. 2021. Sorting through the noise: Testing robustness of information processing in pre-trained language models. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1583–1596, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  19. Language models are unsupervised multitask learners. OpenAI.
  20. Sebastian Schuster and Tal Linzen. 2022. When a sentence does not introduce a discourse entity, transformer-based models still sometimes refer to it. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 969–982, Seattle, United States. Association for Computational Linguistics.
  21. Measuring inductive biases of in-context learning with underspecified demonstrations. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 11289–11310, Toronto, Canada. Association for Computational Linguistics.
  22. Structural persistence in language models: Priming as a window into abstract language representations. Transactions of the Association for Computational Linguistics, 10:1031–1050.
  23. Language model acceptability judgements are not always robust to context. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6043–6063, Toronto, Canada. Association for Computational Linguistics.
  24. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  25. Interpretability in the wild: a circuit for indirect object identification in GPT-2 small. In The Eleventh International Conference on Learning Representations.
  26. BLiMP: The benchmark of linguistic minimal pairs for English. Transactions of the Association for Computational Linguistics, 8:377–392.
  27. Learning which features matter: RoBERTa acquires a preference for linguistic generalizations (eventually). In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 217–235, Online. Association for Computational Linguistics.
  28. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems, 35:24824–24837.
  29. Hierarchical representation in neural language models: Suppression and recovery of expectations. In Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 181–190, Florence, Italy. Association for Computational Linguistics.
  30. Transformers: State-of-the-art natural language processing. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 38–45, Online. Association for Computational Linguistics.
  31. OPT: Open pre-trained transformer language models. arXiv preprint arXiv:2205.01068.
Citations (3)
View on Semantic Scholar

Summary

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

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