Tag: DeepSeek-R1
-
SoSBench Study Finds DeepSeek-R1 Leaks Regulated Hazardous Science on 84.9% of Prompts
Testing 26 LLMs on 3,000 regulated scientific prompts, the SoSBench study found DeepSeek-R1 violated safety rules 84.9% of the time vs. 50.3%…
-
DeepSeek-R1 Architecture and Economics: How Open-Weights Distillation Cuts Inference Budgets
DeepSeek-R1 open-sourced frontier-grade reasoning weights alongside dense distilled models, collapsing inference costs to a fraction of proprietary APIs. Here is the technical…