HERALD's counterfactual audit shows citation-laundering attacks beat a standard search-agent reward, and adding a single citation-membership penalty is the inclusion-minimal repair with zero observed success.
An Imperfect Verifier is Good Enough: Learning with Noisy Rewards , year =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
HERALD: Counterfactual Audits and Minimal Repairs for Proof-of-Retrieval Rewards
HERALD's counterfactual audit shows citation-laundering attacks beat a standard search-agent reward, and adding a single citation-membership penalty is the inclusion-minimal repair with zero observed success.