DynaBench
PulseAugur coverage of DynaBench — every cluster mentioning DynaBench across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
UniGuard framework enhances LLM safety generalization across interfaces
Researchers have developed UniGuard, a novel training framework designed to enhance the safety generalization of guardrail models for LLMs and agentic systems. Unlike existing models that struggle with different interfa…
-
New AdaGuard model adapts to user-defined policies for safer AI agents
Researchers have developed AdaGuard, a novel adaptive guard model designed to enhance the safety of language model agents by allowing user-defined policies. Unlike fixed guard models, AdaGuard can interpret and enforce …
-
New DynaMAC framework enhances robot bimanual manipulation
Researchers have introduced DynaMAC, a new framework designed to improve sample efficiency and generalization in robot manipulation tasks, particularly in dynamic environments and for bimanual coordination. This framewo…
-
New LLM alignment methods explore game theory and configurable safety
Researchers are exploring novel methods for aligning large language models (LLMs) to safety requirements, moving beyond traditional erasure techniques. One approach frames safety as a non-zero-sum game between two LMs, …