Search: (WFRF:(Johansson Moa 1981)) >
The Effect of Scali...
The Effect of Scaling, Retrieval Augmentation and Form on the Factual Consistency of Language Models
-
- Hagström, Lovisa, 1995 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Saynova, Denitsa, 1991 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Norlund, Tobias, 1991 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
show more...
-
- Johansson, Moa, 1981 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Johansson, Richard, 1975 (author)
- Gothenburg University,Göteborgs universitet,Institutionen för data- och informationsteknik (GU),Department of Computer Science and Engineering (GU),University of Gothenburg
-
show less...
-
(creator_code:org_t)
- Association for Computational Linguistics, 2023
- 2023
- English.
-
In: Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 5457–5476, Singapore. - : Association for Computational Linguistics.
- Related links:
-
https://gup.ub.gu.se...
-
show more...
-
https://research.cha...
-
show less...
Abstract
Subject headings
Close
- Large Language Models (LLMs) make natural interfaces to factual knowledge, but their usefulness is limited by their tendency to deliver inconsistent answers to semantically equivalent questions. For example, a model might supply the answer “Edinburgh” to “Anne Redpath passed away in X.” and “London” to “Anne Redpath’s life ended in X.” In this work, we identify potential causes of inconsistency and evaluate the effectiveness of two mitigation strategies: up-scaling and augmenting the LM with a passage retrieval database. Our results on the LLaMA and Atlas models show that both strategies reduce inconsistency but that retrieval augmentation is considerably more efficient. We further consider and disentangle the consistency contributions of different components of Atlas. For all LMs evaluated we find that syntactical form and task artifacts impact consistency. Taken together, our results provide a better understanding of the factors affecting the factual consistency of language models.
Subject headings
- NATURVETENSKAP -- Data- och informationsvetenskap -- Språkteknologi (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences -- Language Technology (hsv//eng)
- HUMANIORA -- Språk och litteratur -- Studier av enskilda språk (hsv//swe)
- HUMANITIES -- Languages and Literature -- Specific Languages (hsv//eng)
Keyword
- language models
- LLMs
- retrieval
- consistency
- accuracy
- evaluation
- knowledge representation
Publication and Content Type
- ref (subject category)
- kon (subject category)
To the university's database