An offline, synthetic benchmark for adversarial evaluation of plan-generation agents, with typed plan traces, response-budget curves, and candidate-frontier diagnostics.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
AdvPlan-Bench: Adversarial Evaluation of Structured Plan-Generation Agents
An offline, synthetic benchmark for adversarial evaluation of plan-generation agents, with typed plan traces, response-budget curves, and candidate-frontier diagnostics.