Follow the Gradient
Filtering the 20% most gradient-aligned examples with SimpleLoRIF reduces GCG attack success rate by up to 22% across Qwen, Llama, and Olmo.
Tracing Safety Degradation Back to Training Data
AI Research Radar
Wednesday, October 28: 2 frontier papers out of 433 read.
Frontier
Papers rated as moving the field
How we chooseTimestral
Timestral foundation models set a new state of the art on GIFT-Eval, TIME, and fev-bench, taking the top two places on every metric.
This week in research
Oct 22–28: 226 cleared, 15 frontier
- Security76◆ 1
- Agents40◆ 3
- Safety39◆ 5
- Robotics19◆ 2
- Society17
- Language11◆ 1
- 4 moreVision, ML, Reasoning, Training10◆ 3
- Inference8
- Science6
- Security76◆ 1
- 8 moreSociety, Language, Inference, Science, Vision, ML, Reasoning, Training52◆ 4
- Agents40◆ 3
- Safety39◆ 5
- Robotics19◆ 2
Size: cleared papers◆ and tint: frontier papers
Models under test
Papers this week that name the model
Labs publishing
Papers this week with an author at the lab
Also cleared
30 papers, by field
Security & privacy14
- Watermark Basins: Black-Box Detection of Distilled LLMsMeta
- Revisiting Privacy Risks in Quantized Deep Neural NetworksCode
- Log-To-Leak: Prompt Injection Attacks on Tool-Using LLM Agents via the Model Context Protocol
- Program-Level Adversarial Co-Adaptive Self-Play for Robotic Security
- SpecGuard: Detecting Multi-Turn LLM Jailbreaks via Specification-Based Runtime Monitoring
All 14 papers in Security & privacy
- Detecting and Mitigating Query-Based System Prompt Leakage Attacks
- Spiking Neural Networks Are Harder to Steal: A Benchmark Study of Model Stealing Attacks
- PHP2Shell and the Future of AI-Discovered Vulnerability Disclosure: Evidence, Verification, and Human Authority
- API-TAIL: Detecting Model Substitution in LLM APIs through Temperature-Aligned Likelihood Tails
- Provenance Signatures for Low Rank Adaptations of Image Diffusion Models
- RePAIR: Inverting Backdoor Attacks in Semantic Segmentation
- Is Your LLM Relay Serving What It Claims? A low-cost and high-power test for LLM Relay Auditing
- Penny for Your Prompts: A Measurement Study of Contextual Leakage in AI Chat Platforms
- Heads of Security: Learned Attention Gates Improve Prompt Injection Robustness
Agents7
- InfraDev: Can Coding Agents Develop AI Infrastructure, and How Can We Tell?Tencent
- Who Convinced Whom? Tracing How LLMs Change Their Minds in Multi-Agent Systems
- Testing Agentic Service Systems: An Agentic Living Lab Methodology Across the Physical-Digital Continuum
- Small agents should know what their tools do
- PROBE: Frontier Coding Agents Find Different Bugs Than Maintainers Fix