AI Capabilities

Applied to Playing Dixit with AI: Can AI Systems Identify Misalignments in My Personalized Statements? by Mariia Koroliuk 8d ago

Applied to o3, Oh My by Tobias D. 9d ago

Applied to INTELLECT-1 Release: The First Globally Trained 10B Parameter Model by Matrice Jacobine 1mo ago

Applied to Have we seen any "ReLU instead of sigmoid-type improvements" recently by KvmanThinking 2mo ago

Applied to A short project on Mamba: grokking & interpretability by Alejandro Tlaie Boria 3mo ago

Applied to The case for a negative alignment tax by Cameron Berg 4mo ago

Applied to On agentic generalist models: we're essentially using existing technology the weakest and worst way you can use it by Yuli_Ban 4mo ago

Applied to Molecular dynamics data will be essential for the next generation of ML protein models by Lauren (often wrong) 4mo ago

Applied to Lifelogging for Alignment & Immortality by Roland 5mo ago

Applied to Diffusion Guided NLP: better steering, mostly a good thing by Nathan Helm-Burger 5mo ago

Applied to "AI achieves silver-medal standard solving International Mathematical Olympiad problems" by Lauren (often wrong) 5mo ago

Applied to How bad would AI progress need to be for us to think general technological progress is also bad? by Jim Buhler 6mo ago

Applied to [Paper] Stress-testing capability elicitation with password-locked models by Lauren (often wrong) 7mo ago

Applied to Memorizing weak examples can elicit strong behavior out of password-locked models by Lauren (often wrong) 7mo ago

Applied to The thing I don't understand about AGI by Jeremy Kalfus 7mo ago

Applied to Getting 50% (SoTA) on ARC-AGI with GPT-4o by Ryan Greenblatt 7mo ago

Applied to What’s the future of AI hardware? by Itay Dreyfus 7mo ago

Applied to [Paper] AI Sandbagging: Language Models can Strategically Underperform on Evaluations by Teun van der Weij 7mo ago

Applied to An Introduction to AI Sandbagging by Teun van der Weij 8mo ago