An Anthropic team has pinpointed locations inside an LLM that map to specific words, people and concepts.
Research ers at Anthropic have mapped portions of the "mind" of one of their AIs, the company reported this week, in what it called "the first ever detailed look inside a modern, production-grade large language model."The new work from Anthropic raises the prospect that generative AI programs like ChatGPT might some day be much easier to understand and control — making them both more useful and, with luck, less dangerous.
The Anthropic project identified "millions" of features in the Claude Sonnet model they studied, but the researchers write that's just a fraction of the whole model. "We don't have an estimate of how many features there are or how we'd know we got all of them ," thesays — and "getting all of them" might require even more computing power than training the model in the first place, an already costly venture.As generative AI becomes easier to directly program, its guardrails might also become more reliable.
Then again, in the wrong hands, the same dials that make the models safer could be used to amp up their capacity for harm.Share on linkedin
Visuals Illustrations Research
France Dernières Nouvelles, France Actualités
Similar News:Vous pouvez également lire des articles d'actualité similaires à celui-ci que nous avons collectés auprès d'autres sources d'information.
Sentiment Analysis through LLM Negotiations: LLM Negotiation for Sentiment AnalysisThis paper introduces a multi-LLM negotiation framework for sentiment analysis.
Lire la suite »
Anthropic Dethroned By Gemini 1.5 Pro’s 1-Million-Token Context WindowGoogle's Gemini 1.5 Pro has achieved a 1-million-token context window, the longest of any large-scale AI foundation model. This expands what these models can accomplish.
Lire la suite »
Anthropic now has a Claude chatbot app for iOSMariella Moon has been a night editor for Engadget since 2013, covering everything from consumer technology and video games to strange little robots that could operate on the human body from the inside one day. She has a special affinity for space, its technologies and its mysteries, though, and has interviewed astronauts for Engadget.
Lire la suite »
Amazon-backed Anthropic launches iPhone app and business tier to compete with OpenAI's ChatGPTAnthropic on Wednesday announced its first-ever enterprise offering, as well as its first iOS app.
Lire la suite »
Anthropic finally releases a Claude mobile appAnthropic releases an iOS app for users to chat using its AI model Claude 3. The company also rolled out a second paid tier focusing on teams.
Lire la suite »
6 Practical Tips for Using Anthropic's Claude ChatbotAnthropic recently launched an iOS app for its Claude chatbot. We asked the company’s head of product design how to get the most out of the AI helper.
Lire la suite »