May 26 – June 1: 12 frontier papers out of 2,217 read.
Papers rated as moving the field
How we choose“Extensive evaluations of leading foundation models and commercial APIs reveal significant performance degradation in these challenging settings, exposing critical evaluation blind spots.”
The authors, in the abstract
A Real-World Multilingual Speech-to-Text Benchmark
“Our evaluation shows that coding agents with frontier models remain below 40% success on both evaluated engines.”
“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).”
“Extensive evaluations demonstrate that JT-Safe-V2 achieves state-of-the-art performance across both general intelligence and safety benchmarks.”
“Across eleven frontier models, we find that the strongest agents pass only half of the worlds and consistently fail on those where latent structure must be uncovered.”
“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.”
“Our experiments demonstrate amplification across diverse biases: from keyword bias to propaganda (e.g., sexism), brand promotion, and instrumental goal-seeking.”
May 26 – Jun 1: 179 cleared, 12 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
167 papers