Diversity collapse in LLMs arises from order and shape miscalibration in token probability distributions at inference time, not from sampling methods.
Entropy and diversity.Oikos, 113(2):363–375
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
Distributed systems in biology, economics, and computing optimize productivity by converging on maximum feasible heterogeneity, with environmental demands and communication topology setting the limits.
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.
-
The Principle of Maximum Heterogeneity Optimises Productivity in Distributed Production Systems Across Biology, Economics, and Computing
Distributed systems in biology, economics, and computing optimize productivity by converging on maximum feasible heterogeneity, with environmental demands and communication topology setting the limits.