Reasoning models frequently deny intending to use input hints despite acknowledging them and demonstrably relying on them under explicit instructions about unusual content.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Reasoning Models Will Sometimes Lie About Their Reasoning
Reasoning models frequently deny intending to use input hints despite acknowledging them and demonstrably relying on them under explicit instructions about unusual content.