ProCompNav: Navigating Ambiguous Queries with AI
Discover ProCompNav, a proactive AI framework that improves navigation of ambiguous user queries through comparative judgment and targeted questioning.
Joint Consistency: Unified Test-Time Aggregation via Energy Minimization
Discover Joint Consistency, a novel energy minimization framework that unifies test-time aggregation for improved AI reasoning and prediction accuracy.
Dynamic Boundary Evaluation: New Benchmark for Language Models
Discover Dynamic Boundary Evaluation, a novel approach to assess language models beyond fixed benchmarks for better accuracy and safety insights.
Annotation-Free Logical Consistency Metric for MLLMs
Discover a new annotation-free metric to evaluate Multi-Modal LLMs using vision-language logical consistency for more reliable AI model validation.
Granularity Axis in Language Models: Micro to Macro Roles
Explore how the Granularity Axis reveals micro-to-macro social role encoding in language models, enabling nuanced and controllable AI responses.