Records found: 5
'MLPerf Inference v5.1 adds DeepSeek-R1, Llama-3.1-8B and Whisper Large V3 while tightening interactive LLM limits; use Closed, scenario and accuracy filters to map results to real SLAs.'
Agentic-R1, developed using the DualDistill framework, integrates natural language reasoning with tool use, outperforming traditional AI models in mathematical problem solving with improved accuracy and efficiency.
Apple researchers have uncovered fundamental weaknesses in large reasoning AI models through controlled puzzle evaluations, showing significant performance drops as task complexity increases.
USC researchers introduce Tina, a family of compact reasoning models that leverage LoRA and reinforcement learning to deliver strong multi-step reasoning performance at a fraction of typical training costs.
Researchers propose DEER, a novel training-free approach allowing large reasoning language models to dynamically exit reasoning early, reducing computation and improving accuracy.