An empirical study claims LLM recommender outputs are sensitive to candidate order and that a new prompting strategy, RISE, reduces that sensitivity.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Evaluating Position Bias in Large Language Model Recommendations
An empirical study claims LLM recommender outputs are sensitive to candidate order and that a new prompting strategy, RISE, reduces that sensitivity.