Thursday, May 28: 3 frontier papers out of 370 read.
Papers rated as moving the field
How we choose“Across three watermarking schemes and three open-source LLMs, the attack triggers 0/6 state-of-the-art content-side statistical detectors while inflating the watermark z-score up to 2.42x (system-level defenses such as entropy-source attestation remain orthogonal and complementary).”
The authors, in the abstract
An Undetectable Integrity-Preserving Attack Against LLM Watermarking
“Experiments on seven strong open-source and closed-source LLMs show that state-of-the-art LLM agents remain vulnerable to Sleeper Attack, even when they achieve low attack success rates under a single-interaction baseline.”
“Experiments on multi-layer transformers confirm the theory and visualize how reasoning is progressively absorbed into deeper layers.”
May 22–28: 121 cleared, 10 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
30 papers