Forced-binary next-token probabilities from an instruction-tuned LLM were saturated (near-deterministic) for 17/18 question pairs, so QQ-equality verdicts did not identify a response mechanism; saturation screening should precede structural interpretation.
World Scientific, Singapore, 2023
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Auditing Question-Order Effects in Large Language Models with the QQ Equality: Mechanism Characterization and a Saturation Caveat
Forced-binary next-token probabilities from an instruction-tuned LLM were saturated (near-deterministic) for 17/18 question pairs, so QQ-equality verdicts did not identify a response mechanism; saturation screening should precede structural interpretation.