Nuovi modi di corrompere i LLM

• Tweet: https://x.com/OwainEvans_UK/status/1999172920506269783
• Paper: https://arxiv.org/pdf/2512.09742

Gemini 3 Flash Preview

• https://blog.google/products/gemini/gemini-3-flash/
• AA: https://artificialanalysis.ai/evaluations/artificial-analysis-intelligence-index (tanti token, poco costo)
• AA: https://artificialanalysis.ai/evaluations/omniscience

GPT Images 1.5

• Blog post: https://openai.com/index/new-chatgpt-images-is-here/
• Lmao: https://x.com/petergostev/status/2001027573636088184; https://lmarena.ai/leaderboard
• Il training data? https://x.com/elder_plinius/status/2001431081170669577