OpenAI, Anthropic AI Research Reveals More About How LLMs Affect Security and Bias

OpenAI, Anthropic AI Research Reveals More About How LLMs Affect Security and Bias

Because large language models operate using neuron-like structures that may link many different concepts and modalities together, it can be difficult for AI developers to adjust their models to change the models’ behavior. If you don’t know what neurons connect what concepts, you won’t know which neurons to change. On May 21, Anthropic published a remarkably detailed map of the inner workings of the fine-tuned version of its Claude AI, specifically the Claude 3 Sonnet…

Read More

Some Generative AI Company Employees Pen Letter Wanting ‘Right to Warn’ About Risks

Some Generative AI Company Employees Pen Letter Wanting ‘Right to Warn’ About Risks

Some current and former employees of OpenAI, Google DeepMind and Anthropic published a letter on June 4 asking for whistleblower protections, more open dialogue about risks and “a culture of open criticism” in the major generative AI companies. The Right to Warn letter illuminates some of the inner workings of the few high-profile companies that sit in the generative AI spotlight. OpenAI holds a distinct status as a nonprofit trying to “navigate massive risks” of…

Read More

Anthropic’s Generative AI Research Reveals More About How LLMs Affect Security and Bias

Anthropic’s Generative AI Research Reveals More About How LLMs Affect Security and Bias

Because large language models operate using neuron-like structures that may link many different concepts and modalities together, it can be difficult for AI developers to adjust their models to change the models’ behavior. If you don’t know what neurons connect what concepts, you won’t know which neurons to change. On May 21, Anthropic created a remarkably detailed map of the inner workings of the fine-tuned version of its Claude 3 Sonnet 3.0 model. With this…

Read More

NIST Establishes AI Safety Consortium

NIST Establishes AI Safety Consortium

Image: Adobe/Grandbrothers The National Institute of Standards and Technology established the AI Safety Institute on Feb. 7 to determine guidelines and standards for AI measurement and policy. U.S. AI companies and companies that do business in the U.S. will be affected by those guidelines and standards and may have the opportunity to have input about them. What is the U.S. AI Safety Institute consortium? The U.S. AI Safety Institute is a joint public and private…

Read More