TechCrunch · AI·9d agoPrismML hopes its tiny LLM will change how we all use AI#bonsai-2#llm-compression#on-device-ai 3 min1
arXiv cs.AI / cs.LG / cs.CL·11d agoWhat Breaks Under Pruning in Smart Homes, and When? Evaluating LLM Degradation Across Architectures and Task Complexity#llm-compression#model-degradation#moeAI research
