Abstract
The field of Math-NLP has witnessed significant growth in recent years, motivated by the desire to expand LLM performance to the leaning of non-linguistic notions (numerals, and subsequently, arithmetic reasoning). However, non-linguistic skill injection typically comes at a cost for LLMs: it leads to catastrophic forgetting of core linguistic skills, a consequence that often remains unaddressed in the literature. As Math-NLP has been able to create LLMs that can closely approximate the mathematical skills of a grade schooler or the arithmetic reasoning skills of a calculator, the practicality of these models fail if they concomitantly shed their linguistic capabilities. In this work, we take a closer look into the phenomena of catastrophic forgetting as it pertains to LLMs and subsequently offer a novel framework for non-linguistic skill injection for LLMs based on information-theoretic interventions and skill-specific losses that enable the learning of strict arithmetic reasoning. Our model outperforms the state-of-the-art both on injected non-linguistic skills and on linguistic knowledge retention, and does so with a fraction of the non-linguistic training data (1/4) and zero additional synthetic linguistic training data.- Anthology ID:
- 2023.acl-long.340
- Volume:
- Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Month:
- July
- Year:
- 2023
- Address:
- Toronto, Canada
- Editors:
- Anna Rogers, Jordan Boyd-Graber, Naoaki Okazaki
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 6178–6191
- Language:
- URL:
- https://aclanthology.org/2023.acl-long.340
- DOI:
- 10.18653/v1/2023.acl-long.340
- Cite (ACL):
- Mandar Sharma, Nikhil Muralidhar, and Naren Ramakrishnan. 2023. Learning Non-linguistic Skills without Sacrificing Linguistic Proficiency. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6178–6191, Toronto, Canada. Association for Computational Linguistics.
- Cite (Informal):
- Learning Non-linguistic Skills without Sacrificing Linguistic Proficiency (Sharma et al., ACL 2023)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-2/2023.acl-long.340.pdf