Submitting one solution per paraphrased variant improves Pass@k over repeating solutions to the original prompt, under a uniform-spread model and in APPS experiments.
Robustness of learning from task instructions, 2022
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Leveraging LLM Inconsistency to Boost Pass@k Performance
Submitting one solution per paraphrased variant improves Pass@k over repeating solutions to the original prompt, under a uniform-spread model and in APPS experiments.