Humans and language models diverge when predicting repeating text
- URL: http://arxiv.org/abs/2310.06408v2
- Date: Mon, 23 Oct 2023 03:15:46 GMT
- Title: Humans and language models diverge when predicting repeating text
- Authors: Aditya R. Vaidya, Javier Turek, Alexander G. Huth
- Abstract summary: We present a scenario in which the performance of humans and LMs diverges.
Human and GPT-2 LM predictions are strongly aligned in the first presentation of a text span, but their performance quickly diverges when memory begins to play a role.
We hope that this scenario will spur future work in bringing LMs closer to human behavior.
- Score: 52.03471802608112
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Language models that are trained on the next-word prediction task have been
shown to accurately model human behavior in word prediction and reading speed.
In contrast with these findings, we present a scenario in which the performance
of humans and LMs diverges. We collected a dataset of human next-word
predictions for five stimuli that are formed by repeating spans of text. Human
and GPT-2 LM predictions are strongly aligned in the first presentation of a
text span, but their performance quickly diverges when memory (or in-context
learning) begins to play a role. We traced the cause of this divergence to
specific attention heads in a middle layer. Adding a power-law recency bias to
these attention heads yielded a model that performs much more similarly to
humans. We hope that this scenario will spur future work in bringing LMs closer
to human behavior.
Related papers
- Reverse-Engineering the Reader [43.26660964074272]
We introduce a novel alignment technique in which we fine-tune a language model to implicitly optimize the parameters of a linear regressor.
Using words as a test case, we evaluate our technique across multiple model sizes and datasets.
We find an inverse relationship between psychometric power and a model's performance on downstream NLP tasks as well as its perplexity on held-out test data.
arXiv Detail & Related papers (2024-10-16T23:05:01Z) - Look Hear: Gaze Prediction for Speech-directed Human Attention [49.81718760025951]
Our study focuses on the incremental prediction of attention as a person is seeing an image and hearing a referring expression.
We developed the Attention in Referral Transformer model or ART, which predicts the human fixations spurred by each word in a referring expression.
In our quantitative and qualitative analyses, ART not only outperforms existing methods in scanpath prediction, but also appears to capture several human attention patterns.
arXiv Detail & Related papers (2024-07-28T22:35:08Z) - Why Does Surprisal From Larger Transformer-Based Language Models Provide
a Poorer Fit to Human Reading Times? [9.909170013118775]
The propensity of larger Transformer-based models to'memorize' sequences during training makes their surprisal estimates diverge from humanlike expectations.
These results suggest that the propensity of larger Transformer-based models to'memorize' sequences during training makes their surprisal estimates diverge from humanlike expectations.
arXiv Detail & Related papers (2022-12-23T03:57:54Z) - Collateral facilitation in humans and language models [0.6091702876917281]
We show that humans display a similar processing advantage for highly anomalous words.
We discuss the implications for our understanding of both human language comprehension and the predictions made by language models.
arXiv Detail & Related papers (2022-11-09T21:08:08Z) - Learning Theory of Mind via Dynamic Traits Attribution [59.9781556714202]
We propose a new neural ToM architecture that learns to generate a latent trait vector of an actor from the past trajectories.
This trait vector then multiplicatively modulates the prediction mechanism via a fast weights' scheme in the prediction neural network.
We empirically show that the fast weights provide a good inductive bias to model the character traits of agents and hence improves mindreading ability.
arXiv Detail & Related papers (2022-04-17T11:21:18Z) - A Generative Language Model for Few-shot Aspect-Based Sentiment Analysis [90.24921443175514]
We focus on aspect-based sentiment analysis, which involves extracting aspect term, category, and predicting their corresponding polarities.
We propose to reformulate the extraction and prediction tasks into the sequence generation task, using a generative language model with unidirectional attention.
Our approach outperforms the previous state-of-the-art (based on BERT) on average performance by a large margins in few-shot and full-shot settings.
arXiv Detail & Related papers (2022-04-11T18:31:53Z) - Better Language Model with Hypernym Class Prediction [101.8517004687825]
Class-based language models (LMs) have been long devised to address context sparsity in $n$-gram LMs.
In this study, we revisit this approach in the context of neural LMs.
arXiv Detail & Related papers (2022-03-21T01:16:44Z) - Multilingual Language Models Predict Human Reading Behavior [8.830621849672108]
We compare the performance of language-specific and multilingual pretrained transformer models to predict reading time measures.
We find that BERT and XLM models successfully predict a range of eye tracking features.
In a series of experiments, we analyze the cross-domain and cross-language abilities of these models and show how they reflect human sentence processing.
arXiv Detail & Related papers (2021-04-12T13:03:49Z) - Adversarial Generative Grammars for Human Activity Prediction [141.43526239537502]
We propose an adversarial generative grammar model for future prediction.
Our grammar is designed so that it can learn production rules from the data distribution.
Being able to select multiple production rules during inference leads to different predicted outcomes.
arXiv Detail & Related papers (2020-08-11T17:47:53Z)
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.