A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Unknown
Introduces Diffusion Transformers (DiTs), a transformer-based architecture for diffusion models that scales effectively with model size and compute.
Unknown
This paper presents new efficient protocols for privacy-preserving machine learning for linear regression, logistic regression, and neural network training using stochastic gradient descent.
Unknown
This paper proposes a recursive self-critiquing framework for scalable oversight of superhuman AI, showing promising results in experiments.
Unknown
This paper proposes collaborative multi-agent debate for scalable oversight, improving error detection in LLM responses.
Unknown
Introduces a principled benchmark for empirically evaluating scalable oversight protocols, enabling competitive comparison of methods for supervising AI systems.
Unknown
Introduces a principled benchmark for empirically studying and competitively evaluating scalable oversight protocols.
Unknown
This paper formalizes scalable oversight in hierarchical reinforcement learning, studying how to scale human feedback to complex tasks.
Unknown
This paper outlines a research paradigm for scalable oversight of large language models, based on Cotra's sandwiching approach, to improve model reliability.
Unknown
This paper explores scalable oversight protocols, specifically evaluating whether weak LLMs can effectively judge strong LLMs, and expresses optimism about debate as a scalable oversight method.
Unknown
This paper investigates scaling laws for scalable oversight, where weaker AI systems monitor stronger ones, to understand how oversight effectiveness scales with model capabilities.
Unknown
This paper introduces a scalable pairwise meta-evaluator to diagnose bias and instability in LLM evaluation, providing a multifaceted view of evaluation dynamics.
Unknown
A scalable algorithm for fast model editing that leverages the low-rank structure of fine-tuning to edit very large pre-trained language models.