Two former OpenAI staffers, Thomas Dimson and Joey Flynn, launched “In the Weights,” a site that measures how readily leading AI models can recall a person without web search, underscoring the shift in reputation discovery from traditional search engines to chatbots. The tool queries models including Grok, Gemini, GPT, Claude and Llama, clusters their responses and assigns a comparative “strength score,” while flagging hallucinations and inconsistencies. Early reception has been strong; the founders plan deeper analyses of why models diverge, where biases appear and which overlooked individuals might warrant Wikipedia entries.
Related articles:
– Large language model
– Hallucination (artificial intelligence)
– Introducing Meta Llama 3: The most capable open LLM to date
– The Claude 3 model family
– Perplexity AI
