How and where does CLIP process negation?
- URL: http://arxiv.org/abs/2407.10488v1
- Date: Mon, 15 Jul 2024 07:20:06 GMT
- Title: How and where does CLIP process negation?
- Authors: Vincent Quantmeyer, Pablo Mosteiro, Albert Gatt,
- Abstract summary: We build on the existence task from the VALSE benchmark to test models' understanding of negation.
We take inspiration from the literature on model interpretability to explain the behaviour of VL models on the understanding of negation.
- Score: 2.5600000778964294
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: 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 particularly interesting issue for multimodal models. However, while such VL benchmarks are useful for measuring model performance, they do not reveal anything about the internal processes 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 negation. Specifically, we approach these questions through an in-depth analysis of the text encoder in CLIP (Radford et al, 2021), a highly influential VL model. We localise parts of the encoder that process negation and analyse the role of attention 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 multimodal 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.
Related papers
- Vision-Language Models Do Not Understand Negation [50.27667000027403]
NegBench is a benchmark designed to evaluate negation understanding across 18 task variations and 79k examples spanning image, video, and medical datasets.
We show that this approach can result in a 10% increase in recall on negated queries and a 40% boost in accuracy on multiple-choice questions with negated captions.
arXiv Detail & Related papers (2025-01-16T09:55:42Z) - LatentQA: Teaching LLMs to Decode Activations Into Natural Language [72.87064562349742]
We introduce LatentQA, the task of answering open-ended questions about model activations in natural language.
We propose Latent Interpretation Tuning (LIT), which finetunes a decoder LLM on a dataset of activations and associated question-answer pairs.
Our decoder also specifies a differentiable loss that we use to control models, such as debiasing models on stereotyped sentences and controlling the sentiment of generations.
arXiv Detail & Related papers (2024-12-11T18:59:33Z) - CV-Probes: Studying the interplay of lexical and world knowledge in visually grounded verb understanding [2.524887615873207]
This study investigates the ability of various vision-language (VL) models to ground context-dependent verb phrases.
We introduce the CV-Probes dataset, containing image-caption pairs with context-dependent verbs.
We employ the MM-SHAP evaluation to assess the contribution of verb tokens towards model predictions.
arXiv Detail & Related papers (2024-09-02T17:39:26Z) - Evaluating Generative Language Models in Information Extraction as Subjective Question Correction [49.729908337372436]
We propose a new evaluation method, SQC-Score.
Inspired by the principles in subjective question correction, we propose a new evaluation method, SQC-Score.
Results on three information extraction tasks show that SQC-Score is more preferred by human annotators than the baseline metrics.
arXiv Detail & Related papers (2024-04-04T15:36:53Z) - Can Large Language Models Understand Context? [17.196362853457412]
This paper introduces a context understanding benchmark by adapting existing datasets to suit the evaluation of generative models.
Experimental results indicate that pre-trained dense models struggle with understanding more nuanced contextual features when compared to state-of-the-art fine-tuned models.
As LLM compression holds growing significance in both research and real-world applications, we assess the context understanding of quantized models under in-context-learning settings.
arXiv Detail & Related papers (2024-02-01T18:55:29Z) - L2CEval: Evaluating Language-to-Code Generation Capabilities of Large
Language Models [102.00201523306986]
We present L2CEval, a systematic evaluation of the language-to-code generation capabilities of large language models (LLMs)
We analyze the factors that potentially affect their performance, such as model size, pretraining data, instruction tuning, and different prompting methods.
In addition to assessing model performance, we measure confidence calibration for the models and conduct human evaluations of the output programs.
arXiv Detail & Related papers (2023-09-29T17:57:00Z) - Language models are not naysayers: An analysis of language models on
negation benchmarks [58.32362243122714]
We evaluate the ability of current-generation auto-regressive language models to handle negation.
We show that LLMs have several limitations including insensitivity to the presence of negation, an inability to capture the lexical semantics of negation, and a failure to reason under negation.
arXiv Detail & Related papers (2023-06-14T01:16:37Z) - Scalable Performance Analysis for Vision-Language Models [26.45624201546282]
Joint vision-language models have shown great performance over a diverse set of tasks.
Our paper introduces a more scalable solution that relies on already annotated benchmarks.
We confirm previous findings that CLIP behaves like a bag of words model and performs better with nouns and verbs.
arXiv Detail & Related papers (2023-05-30T06:40:08Z) - Infusing Finetuning with Semantic Dependencies [62.37697048781823]
We show that, unlike syntax, semantics is not brought to the surface by today's pretrained models.
We then use convolutional graph encoders to explicitly incorporate semantic parses into task-specific finetuning.
arXiv Detail & Related papers (2020-12-10T01:27:24Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.