Greg Cook
Greg Cook reposted
Issues with interpreting p-values haunts even AI, which is prone to same biases as human researchers. ChatGPT, Gemini & Claude all fall prey to "dichotomania" - treating p=0.049 & p=0.051 as categorically different, and paying too much attention to significance. www.cambridge.org/core/journal...
Greg Cook reposted
Greg Cook reposted
As AI models grow in size, their values seem to increasingly converge on the same preferences, whether the model is made by OpenAI or Musk’s X or China’s DeepSeek Not everyone will like resulting AI preferences, so value engineering is likely to become a topic of discussion arxiv.org/abs/2502.08640
Greg Cook reposted
Greg Cook reposted
Thoughtful paper on the topic of “hallucination” metaphors in LLMs from @xolotl.org & @annamillsoer.bsky.social papers.ssrn.com/sol3/papers....
Greg Cook reposted
Greg Cook reposted
Greg Cook reposted
Greg Cook reposted