GUARD is an LLM decoding method combining smoothed global entropy with local entropy spikes to self-tune contrastive search, plus a token-count penalty for speed.
in recognition of distinguished service in the Borneo Territories
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
GUARD: Glocal Uncertainty-Aware Robust Decoding for Effective and Efficient Open-Ended Text Generation
GUARD is an LLM decoding method combining smoothed global entropy with local entropy spikes to self-tune contrastive search, plus a token-count penalty for speed.