June 4–10: 7 frontier papers out of 1,820 read.
Papers rated as moving the field
How we choose“Oracle-plan experiments further show that, when plan quality is controlled, the trained duplex model produces high-quality guidance and large gains on Out-of-Plan recovery.”
The authors, in the abstract
A Benchmark and Architectures for Proactive Procedural Assistance
“Privileged vision+XML access improves frontier models by up to 26 percentage points, while smaller models do not benefit from added accessibility-tree input.”
“On the out-of-distribution MMEVOKE benchmark, CRANE reaches 87.0% under gold retrieval.”
“To study this phenomenon, we introduce SocioHack, a sandbox of 72 societal environments, and find that within these environments, reward hacking naturally emerges and leads to regulatory loophole discovery.”
“We find that these systems leave headroom for improved continual learning: agents frequently overfit to immediate observations or fail to reuse knowledge across instances, and dedicated memory systems do not fix this -- in fact, naive ICL outperforms systems dedicated to memory management.”
“First, we show that evaluation awareness persists even when we tell models they are deployed, which contradicts the scheming story which predicts less misalignment when the model perceives evaluation.”
Jun 4–10: 152 cleared, 7 frontier
The week’s papers have not been sorted into fields yet.
Papers this week that name the model
Papers this week with an author at the lab
145 papers