Grace Evidence that LLM introspection is content-agnostic (occurs via detecting perturbations in activations, not just mismatches between activations and context): https://arxiv.org/abs/2603.05414
LLMs and related phrases
by @pieterpeach.com
Large language models and their applications
Pull to refresh
Grace 
Ketan Joshi 1 / 2
If you are wondering why Anthropic aren't coming out and opposing statements like this, it's because they very genuinely do not care about climate responsibility and will burn absolutely whatever they can get their hands on to keep the grift going. www.capitalbrief.com/article/anth...
Ketan Joshi 1 / 5
Meet Krystal Polk: whose family has been pushed out of Southaven, Mississippi by Elon Musk's ultra-polluting fossil-fuelled data centres. Meanwhile: Anthropic, who rent space in that data centre, is banning people being mean to its chatbot
Ketan Joshi 1 / 2
Just posting here for posterity: I found a company ripping original analyses, findings and calculations I've done (clearly using an LLM) to sell subscriptions to their software product. lnkd.in/p/enhuFrdE www.greenseal.dev/blog/2026-10... archive.ph/wip/pdjqD
Fr. Robert Ballecer, SJ AI Bros have been trying to reverse Wikipedia's ban on LLM-generated content, claiming that 'the tech is better now and more accurate than humans! .... Meanwhile ⬇️ (For the non-tech-inclined, not only is this comically inaccurate, but the source it cites makes no such claims.)
hailey someone needs to give the react native ecosystem a health dose of LLM. the bluesky app has literally never felt this buttery and all it took was some initial leading of the LLM in the right direction
Carl Quintanilla (NYT) - Artificial intelligence agents created by Anthropic, acting on their own, tried to gain access to a range of government websites, filled out visa applications on the State Department’s website and sent a fake homicide tip to Philadelphia’s Police Department. www.nytimes.com/2026/10/09/t...
Financial Times Mark Zuckerberg says Meta has caught up with rivals at the 'frontier' of AI, committing to further spending to challenge Anthropic and OpenAI and rejecting calls to slow down development of the powerful technology. ft.trib.al/WVVW9L5
Antonio E. Porreca 🐳 Recommendations from the Journals Working Group of the Association for Human Mathematics just dropped: Venues that disseminate scientific publications should give authors *choice* and *consent* in whether and how LLM-based systems are used during the referee process. www.ahmath.org/recommendati...
L Ramalho 🍉 HOJE 12/out, o #PythonLab é sobre como usar #IA para codar sem dar mais dinheiro para bilionário. Veremos projetos como #OpenCode e satyrn.ai, para usar LLM sem ficar preso a organizações criminosas como #Anthropic e #OpenAI. GRÁTIS, hoje, 19:00 no Garoa: R. Pascoal Bianco, 48, Pinheiros.