surprisal

Tag

Cards List
#surprisal

Do Cantonese-Adapted Language Models Better Predict Cantonese Reading? A Cross-Model Eye-Tracking Evaluation

arXiv cs.CL · 2026-09-03 Cached

This paper evaluates whether Cantonese-adapted language models better predict Cantonese reading using eye-tracking data, comparing models like CKIP GPT-2 and CantoneseLLM-7B. Results indicate that more extensive Cantonese-specific training improves predictive fit, though performance varies by information-theoretic measure.

0 favorites 0 likes
#surprisal

Human-Like Anaphor Resolution in Large Language Models

arXiv cs.CL · 2026-08-07 Cached

This paper investigates whether five open-weight LLMs exhibit human-like sensitivity to psycholinguistic factors in anaphor resolution, using surprisal and comprehension accuracy as behavioral measures. Results show selective cognitive alignment, with some models matching human discourse sensitivity but not semantic interference effects.

0 favorites 0 likes
#surprisal

Encoding EEG Signals to Examine Human-Like Next-Word Prediction Behaviour in Language Models

arXiv cs.CL · 2026-07-21 Cached

This paper investigates whether language models' next-word prediction aligns with human cognitive processing by analyzing EEG signals and event-related potentials, finding that only surprisal correlates with human brain responses, especially for open-class words.

0 favorites 0 likes
#surprisal

Contextual Semantic Relevance Tracks fMRI BOLD Responses During Naturalistic Speech Comprehension

arXiv cs.CL · 2026-07-20 Cached

This study demonstrates that contextual semantic relevance, measuring how strongly an incoming word relates to its recent semantic context, reliably predicts fMRI BOLD responses during naturalistic speech comprehension across two datasets, whereas surprisal (local probabilistic expectation) does not. The findings support that slow hemodynamic responses are especially sensitive to contextual semantic integration rather than local prediction.

0 favorites 0 likes
#surprisal

STARE: Surprisal-Guided Token-Level Advantage Reweighting for Policy Entropy Stability

Hugging Face Daily Papers · 2026-06-17 Cached

STARE addresses policy entropy collapse in GRPO-based reinforcement learning for large language models by introducing surprisal-guided token-level advantage reweighting and target-entropy regulation, achieving 4%-8% accuracy gains on AIME benchmarks.

0 favorites 0 likes
#surprisal

Trajectory Dynamics in Language Model Hidden States Predict Human Processing Costs Beyond Surprisal

arXiv cs.CL · 2026-06-05 Cached

Introduces trajectory extrapolation error, a measure derived from transformer LM hidden states that predicts human reading times independently of and orthogonally to surprisal, revealing a dissociable component of incremental processing cost.

0 favorites 0 likes
#surprisal

Why are language models less surprised than humans? Testing the Parse Multiplicity Mismatch Hypothesis

arXiv cs.CL · 2026-05-18 Cached

This paper tests the Parse Multiplicity Mismatch Hypothesis, proposing that language models underpredict human processing difficulty in garden path sentences because they can consider more simultaneous parses. Using RNNGs with beam search, they find reducing the number of active parses increases predicted garden path effects, but not enough to fully capture human data.

0 favorites 0 likes
← Back to home

Submit Feedback