Block-based double decoders use doubly-causal block attention masks to combine decoder-only training efficiency with encoder-decoder inference efficiency, outperforming standard encoder-decoders in scaling experiments.
Gomez, Łukasz Kaiser, and Illia Polosukhin
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
LLMs encode accurate but brittle internal beliefs about latent game states and convert them poorly into actions, creating systematic gaps that explain strategic failures.
citing papers explorer
-
Block-Based Double Decoders
Block-based double decoders use doubly-causal block attention masks to combine decoder-only training efficiency with encoder-decoder inference efficiency, outperforming standard encoder-decoders in scaling experiments.
-
Why Do LLMs Struggle in Strategic Play? Broken Links Between Observations, Beliefs, and Actions
LLMs encode accurate but brittle internal beliefs about latent game states and convert them poorly into actions, creating systematic gaps that explain strategic failures.