Sélection automatique · sans validation scientifique par ce site
Langue originale : anglais
Evaluation & safety · provider_announcement
Alongside other leading AI companies, we're committed to robust child safety measures in the development and deployment of generative AI.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Claude 3 Haiku is the fastest and most affordable model in its intelligence class.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Interpretability · provider_announcement
Why qualitative research plays a central role in interpretability, with heuristics for judging the quality of qualitative work.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Evaluation & safety · provider_announcement
The Claude 3 model family—Haiku, Sonnet, and Opus—sets new industry benchmarks across a wide range of cognitive tasks.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Agent systems · provider_announcement
Claude 2.1 brings a 200K token context window, a 2x reduction in hallucination rates, beta tool use, and system prompts.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Amazon will invest up to $4 billion in Anthropic and become its primary cloud provider, with models trained on AWS Trainium and Inferentia.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Our Responsible Scaling Policy is a series of technical and organizational protocols to manage the risks of increasingly capable AI systems.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Anthropic and Boston Consulting Group partner to bring Claude to BCG's enterprise clients for knowledge management and data analysis.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Claude Pro is a subscription for claude.ai offering 5x more usage, priority access during peak periods, and early access to new features.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Mathematical reasoning · provider_announcement
Breaking complex questions into subquestions produces more faithful reasoning than chain-of-thought while retaining accuracy gains.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Mathematical reasoning · provider_announcement
Does chain-of-thought reasoning reflect a model's actual decision process? We find faithfulness varies by task and declines in larger models.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Evaluation & safety · provider_announcement
Anthropic's NTIA submission recommends rigorous model evaluations, pre-registration of large training runs, and mandatory external red teaming.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Mathematical reasoning · provider_announcement
We’ve trained a model to achieve a new state-of-the-art in mathematical problem solving by rewarding each correct step of reasoning (“process supervision”) instead of simply
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Complexity theory · provider_announcement
Developing interpretability ideas on dictionary learning, defining features, and the possibility of superposition in attention heads.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Claude's context window expands from 9K to 100K tokens, roughly 75,000 words, letting it process hundreds of pages in under a minute.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
How Claude's constitution—explicit principles drawn from sources like the UN Declaration of Human Rights—guides its training.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Anthropic urges Congress to ambitiously fund NIST to build AI measurement tools, standards, and testbeds.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Evaluation & safety · provider_announcement
Transformative AI may arrive within a decade, and we don't yet know how to make it safe. We outline the research directions we're pursuing to
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Compute & efficiency · provider_announcement
Anthropic selects Google Cloud as preferred cloud provider, using its GPU and TPU clusters to train, scale, and deploy AI systems.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Evaluation & safety · provider_announcement
Anthropic raised $580 million in Series B funding to build infrastructure for studying and improving the safety of large-scale AI models.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Formal verification · provider_announcement
We built a neural theorem prover for Lean that learned to solve a variety of challenging high-school olympiad problems, including problems from the AMC12 and
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Complexity theory · provider_announcement
Reverse-engineering small attention-only transformers reveals how induction heads form via head composition.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Mathematical reasoning · provider_announcement
We’ve trained a system that solves grade school math problems with nearly twice the accuracy of a fine-tuned GPT-3 model. It solves about 90% as
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗AI models · provider_announcement
Anthropic's Series A, led by Jaan Tallinn, funds research toward steerable, interpretable, and robust large-scale AI systems.
Sélection automatique · sans validation scientifique par ce siteLire la note de recherche ↗Aucune note correspondante.