Daniel Khashabi

2 articles on SOTA Papers

Million-Clip Dataset Tests Video Reasoning Beyond Visual Quality

VBVR pairs 200 curated reasoning tasks with rule-based, human-aligned evaluation to study whether video models generalize across spatiotemporal reasoning problems.

cs.AI cs.LG cs.MM

Aug 30, 20264 min2602.20159

Computer SciencePick

Neuro-Symbolic Reasoning Improves LLM Accuracy by 12-18%

Generating natural language explanations then verifying them symbolically outperforms chain-of-thought prompting.

cs.AI cs.LG cs.CL

Aug 5, 202410 min2408.01234 Code available