Diversity collapse in LLMs arises from order and shape miscalibration in token probability distributions at inference time, not from sampling methods.
Evaluating the quality of randomness and entropy in tasks supported by large language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Sampling More, Getting Less: Calibration is the Diversity Bottleneck in LLMs
Diversity collapse in LLMs arises from order and shape miscalibration in token probability distributions at inference time, not from sampling methods.