Predict the Next Word: Humans exhibit uncertainty in this task and language models _____

• URL: http://arxiv.org/abs/2402.17527v2
• Date: Mon, 18 Mar 2024 16:21:24 GMT
• Title: Predict the Next Word: Humans exhibit uncertainty in this task and language models _____
• Authors: Evgenia Ilia, Wilker Aziz,
• Abstract summary: Language models (LMs) are trained to assign probability to human-generated text. We exploit this fact and evaluate the LM's ability to reproduce variability that humans exhibit in the 'next word prediction' task. We assess GPT2, BLOOM and ChatGPT and find that they exhibit fairly low calibration to human uncertainty.
• Score: 7.581259361859477
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Language models (LMs) are statistical models trained to assign probability to human-generated text. As such, it is reasonable to question whether they approximate linguistic variability exhibited by humans well. This form of statistical assessment is difficult to perform at the passage level, for it requires acceptability judgements (i.e., human evaluation) or a robust automated proxy (which is non-trivial). At the word level, however, given some context, samples from an LM can be assessed via exact matching against a prerecorded dataset of alternative single-word continuations of the available context. We exploit this fact and evaluate the LM's ability to reproduce variability that humans (in particular, a population of English speakers) exhibit in the 'next word prediction' task. This can be seen as assessing a form of calibration, which, in the context of text classification, Baan et al. (2022) termed calibration to human uncertainty. We assess GPT2, BLOOM and ChatGPT and find that they exhibit fairly low calibration to human uncertainty. We also verify the failure of expected calibration error (ECE) to reflect this, and as such, advise the community against relying on it in this setting.

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)
• Linguistic Calibration of Long-Form Generations [57.836339732160916]
  Language models (LMs) may lead their users to make suboptimal downstream decisions when they confidently hallucinate. This issue can be mitigated by having the LM verbally convey the probability that its claims are correct, but existing models cannot produce long-form text with calibrated confidence statements. We define linguistic calibration for long-form generations: an LM is linguistically calibrated if its generations enable its users to make calibrated probabilistic predictions.
  arXiv  Detail & Related papers  (2024-03-30T20:47:55Z)
• Temperature-scaling surprisal estimates improve fit to human reading times -- but does it do so for the "right reasons"? [15.773775387121097]
  We show that calibration of large language models typically improves with model size. We find that temperature-scaling probabilities lead to a systematically better fit to reading times.
  arXiv  Detail & Related papers  (2023-11-15T19:34:06Z)
• Humans and language models diverge when predicting repeating text [52.03471802608112]
  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.
  arXiv  Detail & Related papers  (2023-10-10T08:24:28Z)
• Towards Fine-Grained Information: Identifying the Type and Location of Translation Errors [80.22825549235556]
  Existing approaches can not synchronously consider error position and type. We build an FG-TED model to predict the textbf addition and textbfomission errors. Experiments show that our model can identify both error type and position concurrently, and gives state-of-the-art results.
  arXiv  Detail & Related papers  (2023-02-17T16:20:33Z)
• Naturalistic Causal Probing for Morpho-Syntax [76.83735391276547]
  We suggest a naturalistic strategy for input-level intervention on real world data in Spanish. Using our approach, we isolate morpho-syntactic features from counfounders in sentences. We apply this methodology to analyze causal effects of gender and number on contextualized representations extracted from pre-trained models.
  arXiv  Detail & Related papers  (2022-05-14T11:47:58Z)
• Evaluating Distributional Distortion in Neural Language Modeling [81.83408583979745]
  A heavy-tail of rare events accounts for a significant amount of the total probability mass of distributions in language. Standard language modeling metrics such as perplexity quantify the performance of language models (LM) in aggregate. We develop a controlled evaluation scheme which uses generative models trained on natural data as artificial languages.
  arXiv  Detail & Related papers  (2022-03-24T01:09:46Z)
• Translation Error Detection as Rationale Extraction [36.616561917049076]
  We study the behaviour of state-of-the-art sentence-level QE models and show that explanations can indeed be used to detect translation errors. We introduce a novel semi-supervised method for word-level QE and (ii) propose to use the QE task as a new benchmark for evaluating the plausibility of feature attribution.
  arXiv  Detail & Related papers  (2021-08-27T09:35:14Z)
• Double Perturbation: On the Robustness of Robustness and Counterfactual Bias Evaluation [109.06060143938052]
  We propose a "double perturbation" framework to uncover model weaknesses beyond the test dataset. We apply this framework to study two perturbation-based approaches that are used to analyze models' robustness and counterfactual bias in English.
  arXiv  Detail & Related papers  (2021-04-12T06:57:36Z)
• Prediction Confidence from Neighbors [0.0]
  The inability of Machine Learning (ML) models to successfully extrapolate correct predictions from out-of-distribution (OoD) samples is a major hindrance to the application of ML in critical applications. We show that feature space distance is a meaningful measure that can provide confidence in predictions. This enables earlier and safer deployment of models in critical applications and is vital for deploying models under ever-changing conditions.
  arXiv  Detail & Related papers  (2020-03-31T09:26:09Z)

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.