Chat UI and API access to the same chatbot produce different accuracy, consistency, citation, and refusal behaviors on safety benchmarks, and web search changes these patterns further.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.HC 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
What Current AI Benchmarks Leave Unmeasured: Modality, Search, Citations, and Implications (for Safety Evaluations)
Chat UI and API access to the same chatbot produce different accuracy, consistency, citation, and refusal behaviors on safety benchmarks, and web search changes these patterns further.