Weak-to-strong generalization
We present a new research direction for superalignment, together with promising initial results: can we leverage the generalization properties of deep learning to control strong models with weak supervisors?
We present a new research direction for superalignment, together with promising initial results: can we leverage the generalization properties of deep learning to control strong models with weak supervisors?
In the News In AI copyright case, Zuckerberg turns to YouTube for his defense Meta CEO Mark Zuckerberg appears to have used YouTube’s battle to remove pirated content to defend his own company’s use of a data set containing copyrighted e-books, reveals newly released snippets of a deposition he gave late last year. techcrunch.com Sponsor…
Model Predictive Control (MPC) has become a key technology in a number of fields, including power systems, robotics, transportation, and process control. Sampling-based MPC has shown effectiveness in applications such as path planning and control, and it is useful as a subroutine in Model-Based Reinforcement Learning (MBRL), all because of its versatility and parallelizability, Despite…
This is 100 Years From Now. Once a week we skip a century and try to picture what life actually looks like when the stuff we’re building now has had time to settle in. This week: the last vote.
Anthropic confidentially filed a draft S-1 with the SEC today for a proposed public offering. The company also shipped Claude Opus 4.8 last week with a 4x code-reliability gain. NVIDIA used GTC Taipei to open Cosmos 3, ramp Vera Rubin into production, and put a 1-petaflop AI box on developer laptops. Google retires Gemini 2.0…
Tensoic has recently introduced Kannada Llama (Kan-LLaMA) to address the limitations of language models (LLMs), focusing specifically on proprietary characteristics, computational resources, and barriers to broader research community contributions. Emphasize the importance of open models using mouth to facilitate innovation in natural language processing (NLP) and machine translation with emphasis. Despite the success of models…
Welcome Things are pretty serious. AI is totally a geopolitical issue now – and more AI companies have 1 billion seed rounds now Anthropic got blacklisted. OpenAI is the bad guy as users revolted. LeCun walked away from Meta with a billion dollars and a thesis that everyone else is wrong. Oracle said it out…