skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Testing Causal Models of Word Meaning in LLMs
Large Language Models (LLMs) have driven extraordinary improvements in NLP. However, it is unclear how such models represent lexical concepts-i.e., the meanings of the words they use. We evaluate the lexical representations of GPT-4, GPT-3, and Falcon-40B through the lens of HIPE theory, a concept representation theory focused on words describing artifacts (such as ‚Äúmop‚Äù, ‚Äúpencil‚Äù, and ‚Äúwhistle‚Äù). The theory posits a causal graph relating the meanings of such words to the form, use, and history of the referred objects. We test LLMs with the stimuli used by Chaigneau et al. (2004) on human subjects, and consider a variety of prompt designs. Our experiments concern judgements about causal outcomes, object function, and object naming. We do not find clear evidence that GPT-3 or Falcon-40B encode HIPE's causal structure, but find evidence that GPT-4 does. The results contribute to a growing body of research characterizing the representational capacity of LLMs.  more » « less
Award ID(s):
1956221
PAR ID:
10559551
Author(s) / Creator(s):
;
Publisher / Repository:
Proceedings of the Annual Meeting of the Cognitive Science Society Volume 46
Date Published:
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; (, Association for Computational Linguistics)
    In-context learning (ICL) is an important paradigm for adapting large language models (LLMs) to new tasks, but the generalization behavior of ICL remains poorly understood. We investigate the inductive biases of ICL from the perspective of feature bias: which feature ICL is more likely to use given a set of underspecified demonstrations in which two features are equally predictive of the labels. First, we characterize the feature biases of GPT-3 models by constructing underspecified demonstrations from a range of NLP datasets and feature combinations. We find that LLMs exhibit clear feature biases—for example, demonstrating a strong bias to predict labels according to sentiment rather than shallow lexical features, like punctuation. Second, we evaluate the effect of different interventions that are designed to impose an inductive bias in favor of a particular feature, such as adding a natural language instruction or using semantically relevant label words. We find that, while many interventions can influence the learner to prefer a particular feature, it can be difficult to overcome strong prior biases. Overall, our results provide a broader picture of the types of features that ICL may be more likely to exploit and how to impose inductive biases that are better aligned with the intended task. 
    more » « less
  2. ; ; ; (, Transactions on machine learning research)
    The causal capabilities of large language models (LLMs) are a matter of significant debate, with critical implications for the use of LLMs in societally impactful domains such as medicine, science, law, and policy. We conduct a "behavorial" study of LLMs to benchmark their capability in generating causal arguments. Across a wide range of tasks, we find that LLMs can generate text corresponding to correct causal arguments with high probability, surpassing the best-performing existing methods. Algorithms based on GPT-3.5 and 4 outperform existing algorithms on a pairwise causal discovery task (97%, 13 points gain), counterfactual reasoning task (92%, 20 points gain) and event causality (86% accuracy in determining necessary and sufficient causes in vignettes). We perform robustness checks across tasks and show that the capabilities cannot be explained by dataset memorization alone, especially since LLMs generalize to novel datasets that were created after the training cutoff date. That said, LLMs exhibit unpredictable failure modes, and we discuss the kinds of errors that may be improved and what are the fundamental limits of LLM-based answers. Overall, by operating on the text metadata, LLMs bring capabilities so far understood to be restricted to humans, such as using collected knowledge to generate causal graphs or identifying background causal context from natural language. As a result, LLMs may be used by human domain experts to save effort in setting up a causal analysis, one of the biggest impediments to the widespread adoption of causal methods. Given that LLMs ignore the actual data, our results also point to a fruitful research direction of developing algorithms that combine LLMs with existing causal techniques. Code and datasets are available at https://github.com/py-why/pywhy-llm. 
    more » « less
  3. ; ; (, Interspeech)
    Training emotion recognition models has relied heavily on human annotated data, which present diversity, quality, and cost challenges. In this paper, we explore the potential of Large Language Models (LLMs), specifically GPT-4, in automating or assisting emotion annotation. We compare GPT-4 with supervised models and/or humans in three aspects: agreement with human annotations, alignment with human perception, and impact on model training. We find that common metrics that use aggregated human annotations as ground truth can underestimate GPT-4's performance, and our human evaluation experiment reveals a consistent preference for GPT-4 annotations over humans across multiple datasets and evaluators. Further, we investigate the impact of using GPT-4 as an annotation filtering process to improve model training. Together, our findings highlight the great potential of LLMs in emotion annotation tasks and underscore the need for refined evaluation methodologies. 
    more » « less
  4. ; ; (, Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers))
    Because meaning can often be inferred from lexical semantics alone, word order is often a redundant cue in natural language. For example, the words chopped, chef, and onion are more likely used to convey “The chef chopped the onion,” not “The onion chopped the chef.” Recent work has shown large language models to be surprisingly word order invariant, but crucially has largely considered natural prototypical inputs, where compositional meaning mostly matches lexical expectations. To overcome this confound, we probe grammatical role representation in English BERT and GPT-2, on instances where lexical expectations are not sufficient, and word order knowledge is necessary for correct classification. Such non-prototypical instances are naturally occurring English sentences with inanimate subjects or animate objects, or sentences where we systematically swap the arguments to make sentences like “The onion chopped the chef”. We find that, while early layer embeddings are largely lexical, word order is in fact crucial in defining the later-layer representations of words in semantically non-prototypical positions. Our experiments isolate the effect of word order on the contextualization process, and highlight how models use context in the uncommon, but critical, instances where it matters. 
    more » « less
  5. ; ; (, Journal of experimental psychology. General)
    It is well-known that children rapidly learn words, following a range of heuristics. What is less well appreciated is that – because most words are polysemous and have multiple meanings (e.g., ‘glass’ can label a material and drinking vessel) – children will often be learning a new meaning for a known word, rather than an entirely new word. Across four experiments we show that children flexibly adapt a well-known heuristic – the shape bias – when learning polysemous words. Consistent with previous studies, we find that children and adults preferentially extend a new object label to other objects of the same shape. But we also find that when a new word for an object (‘a gup’) has previously been used to label the material composing that object (‘some gup’), children and adults override the shape bias, and are more likely to extend the object label by material (Experiments 1 and 3). Further, we find that, just as an older meaning of a polysemous word constrains interpretations of a new word meaning, encountering a new word meaning leads learners to update their interpretations of an older meaning (Experiment 2). Finally, we find that these effects only arise when learners can perceive that a word’s meanings are related, not when they are arbitrarily paired (Experiment 4). Together, these findings show that children can exploit cues from polysemy to infer how new word meanings should be extended, suggesting that polysemy may facilitate word learning and invite children to construe categories in new ways. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email