EdgeBench
PulseAugur coverage of EdgeBench — every cluster mentioning EdgeBench across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
AI model evaluations need richer toolkits beyond simple scores
Current AI model evaluation methods, primarily relying on benchmarks, are insufficient for accurately assessing both capabilities and safety. These benchmarks suffer from issues like saturation, unreliability, and gamea…
-
ByteDance launches Seed STEM Scientist Program to explore AI in scientific discovery
ByteDance's Seed Edge team has launched the Seed STEM Scientist Program, aiming to recruit 100 scholars to explore how AI can accelerate scientific discovery. The program will focus on real-world research problems, prov…
-
ByteDance discovers AI scaling law for faster agent learning · 1 source tracked
ByteDance researchers have identified a new scaling law suggesting that AI agents can double their learning speed every three months through interaction with real-world tasks. This discovery, detailed in a research pape…
-
EdgeBench paper reveals log-sigmoid scaling law for AI learning speed
A new research paper introduces EdgeBench, a framework for studying how AI agents learn from real-world environments after deployment. Analysis of 38,000 hours of agent interaction across 134 tasks reveals a log-sigmoid…