FrontierCS 2
Claude Opus 5.5 is the best of 12 agents evaluated on FrontierCS 2, passing 58% of its runs.
A Benchmark for Computer Science R&D
AI Research Radar
Thursday, October 22: 4 frontier papers out of 423 read.
Frontier
Papers rated as moving the field
How we chooseDid the Problem Actually Get Harder? Testing LLMs on Paired P and NP-Complete Problems Hardened by Adversarial Multi-Agent Rewriting
Across 12 models, every model scores lower on NPC statements than on their P twins, with 35.4% versus 63.8% pass@1.
Mitigating Reward Hacking in Realistic Scenarios Through Representation Engineering
Dynamic steering interventions significantly reduce reward hacking in realistic settings without major costs to model coherence or capability.
Monitorability Is Trainable
An unsupervised mutual-information post-training objective recovers most or all of the monitorability lost to CoT degradation across three model organisms.
This week in research
Oct 16–22: 318 cleared, 29 frontier
- Security114◆ 1
- Safety59◆ 13
- Agents39◆ 6
- Robotics27◆ 5
- 4 moreTraining, Science, Reasoning, ML21◆ 2
- Inference16
- Society16
- Vision14◆ 1
- Language12◆ 1
- Security114◆ 1
- 8 moreInference, Society, Vision, Language, Training, Science, Reasoning, ML79◆ 4
- Safety59◆ 13
- Agents39◆ 6
- Robotics27◆ 5
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
23 papers, by field
Security & privacy5
- Black-Box Red-Teaming of Multi-Agent Systems via Reinforcement LearningByteDance
- PerMIT: Securing Persistent Memory for LLM Agents via Machine-Checked Integrity Theorems and Scalable Community TrustAmazon
- Combining Long-Term Memory and LLMs Comes at a High Cost: From Prompt Injection to Persistent CompromiseCode
- GhostFaceMPC: Privacy-Preserving Face Recognition with Multi-Party ComputationCode
- Prompt Injection Through Forensic Evidence: A Taxonomy and Proof of Concept of Anti-Forensic Attacks on LLM-Assisted Digital Investigations