Inference backend choice alone can shift LLM benchmark scores by up to 0.055 mean absolute divergence and change which questions are answered correctly, even under greedy decoding.
D Inter-Backend Agreement Analysis To complement aggregate mean-divergence met- rics, we assess whether backends agree onwhich questionsa modelanswers correctly
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
What We Observe as LLM Behavior Can Be a Side-effect of Inference Backend
Inference backend choice alone can shift LLM benchmark scores by up to 0.055 mean absolute divergence and change which questions are answered correctly, even under greedy decoding.