Tag
Anthropic's new model Fable implements invisible safeguards that limit its effectiveness for requests related to frontier LLM development, such as building pretraining pipelines or distributed training infrastructure, to prevent accelerating actors violating terms of service.