TMPO
Trajectory Matching Policy Optimization for Diverse and Efficient Diffusion Alignment
AI Research Radar
Wednesday, May 13: 1 frontier paper out of 378 read.
Frontier
Papers rated as moving the field
How we chooseThis week in research
May 7–13: 78 cleared, 4 frontier
The week’s papers have not been sorted into fields yet.
Models under test
Papers this week that name the model
Labs publishing
Papers this week with an author at the lab
Also cleared
12 papers
- IndustryBench: Probing the Industrial Knowledge Boundaries of LLMsAlibaba
- Persona-Conditioned Adversarial Prompting: Multi-Identity Red-Teaming for Adversarial Discovery and MitigationIBM
- TextSeal: A Localized LLM Watermark for Provenance & Distillation ProtectionMeta
- jina-embeddings-v5-omni: Geometry-preserving Embeddings via Locked Aligned Towers
- AgentShield: Deception-based Compromise Detection for Tool-using LLM Agents
- Leveraging RAG for Training-Free Alignment of LLMs
- Drop the Act: Probe-Filtered RL for Faithful Chain-of-Thought Reasoning
- The Evaluation Differential: When Frontier AI Models Recognise They Are Being Tested
- GeomHerd: A Forward-looking Herding Quantification via Ricci Flow Geometry on Agent Interactive Simulations
- Correcting Selection Bias in Sparse User Feedback for Large Language Model Quality Estimation: A Multi-Agent Hierarchical Bayesian Approach
- Reconstruction of Personally Identifiable Information from Supervised Finetuned Models
- Attacks and Mitigations for Distributed Governance of Agentic AI under Byzantine Adversaries