NTP-trained LLMs can predict a second token by marginalizing over first-token probabilities, but adapting them with extra heads only partially closes the gap to that baseline.
Overview of the IWSLT 2017 evaluation campaign
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
-
On multi-token prediction for efficient LLM inference
NTP-trained LLMs can predict a second token by marginalizing over first-token probabilities, but adapting them with extra heads only partially closes the gap to that baseline.