| Title: |
How and where does CLIP process negation? |
| Authors: |
Quantmeyer, Vincent; Mosteiro Romero, Pablo; Gatt, Albert; Leerstoel Schoot; Methodology and Statistics for the Behavioural and Social Sciences; Sub Natural Language Processing; Natural Language Processing |
| Publication Year: |
2024 |
| Subject Terms: |
Language and Linguistics; Computer Science Applications; Software; Ophthalmology; Linguistics and Language |
| Description: |
Various benchmarks have been proposed to test linguistic understanding in pre-trained vision & language (VL) models. Here we build on the existence task from the VALSE benchmark (Parcalabescu et al., 2022) which we use to test models’ understanding of negation, a par- ticularly interesting issue for multimodal mod- els. However, while such VL benchmarks are useful for measuring model performance, they do not reveal anything about the internal pro- cesses through which these models arrive at their outputs in such visio-linguistic tasks. We take inspiration from the growing literature on model interpretability to explain the behaviour of VL models on the understanding of nega- tion. Specifically, we approach these questions through an in-depth analysis of the text encoder in CLIP (Radford et al., 2021), a highly influen- tial VL model. We localise parts of the encoder that process negation and analyse the role of at- tention heads in this task. Our contributions are threefold. We demonstrate how methods from the language model interpretability literature (such as causal tracing) can be translated to mul- timodal models and tasks; we provide concrete insights into how CLIP processes negation on the VALSE existence task; and we highlight inherent limitations in the VALSE dataset as a benchmark for linguistic understanding. |
| Document Type: |
book part |
| File Description: |
application/pdf |
| Language: |
English |
| Relation: |
https://dspace.library.uu.nl/handle/1874/482164 |
| Availability: |
https://dspace.library.uu.nl/handle/1874/482164 |
| Rights: |
info:eu-repo/semantics/OpenAccess |
| Accession Number: |
edsbas.F3C5E680 |
| Database: |
BASE |