Tag
Co-ReAct introduces a rubric-guided action-selection framework that uses rubrics as step-level guidance during inference for ReAct agents, improving trajectory quality and outperforming baselines on DeepResearchBench and SQA-CS-V2.
Proposes VeGAS, a test-time framework for MLLM-based embodied agents that samples multiple candidate actions and uses a generative verifier to select the most reliable, achieving up to 36% relative improvement over CoT baselines on challenging tasks.