Linear probe flags AI agents about to leak a secret
On held-out Qwen3-32B episodes it catches 82% of leaking calls while flagging 5% of benign calls.
48
AI model
GPT-oss-20b
- Mentions · 7d0
- Mentions · 30d3
- All mentions9
- First seenSep 2, 2026
- Last mentioned13d ago
Timeline
9 articles · newest firstHeads of Security attention gates improve prompt-injection robustness
Authors report RL-Hammer attack success of 10% on HoS versus 97% on GPT-OSS-20B and 85% on SecAlign.
52
Fin-SD tests how much hint signal self-distillation needs
On GPT-OSS-20B, a generic hint matched the full 3.03-point gain and used 9.5% fewer tokens than full-solution distillation.
42
Agent Name Collision Attacks in Multi-Agent Systems
A multi-agent system vulnerability is identified where remote names are used as routing identifiers, causing wrong-peer dispatch.
40
CCR: Towards a Common, Quality-Gated CACAO Integrations Registry for European Cybersecurity Automation
A paper introduces a standardized CACAO integration registry with quality-gating for cybersecurity integration management.
10
Control-Token Injection Suppresses Chain-of-Thought and Defeats Reasoning-Based Oversight in Tool-Using Agents
This paper demonstrates that controlling token injection can suppress chain-of-thought in tool-using agents.
40
MDRC: A Deployable State-Recovery Defense for Traffic Signal Control under Sensor Corruption
This paper presents a state-recovery defense for traffic signal control under sensor corruption.
10
Issuer-Sovereign Agentic Payments
This paper describes a method to keep issuing bank control over agent payments.
10
What Else Needs Fixing? Exploring Cost-Effective Test-Time Compute for Revision Propagation in Artifacts Generated Through Conversation
A new benchmark shows LLMs reach 68.3-93% accuracy propagating local revisions across conversationally generated artifacts, with parallel-sample selection most cost-effective.
30
Appears with
- LockBitmalware · 5
- Gemma agentproduct · 5
- Microsoftorganization · 5
- adversariesthreat actor · 5
- gpt-oss-120bai model · 2
- Qwen3-32Bai model · 1
- Llama-3.3-70B-Instructai model · 1
- qwen3.5product · 1
- NTTvendor · 1
- gpt-5.4-miniproduct · 1
- qwen3.5-9bai model · 1
- gpt-ossproduct · 1
- gpt-5.4-miniai model · 1
- qwen3.5-122bai model · 1
- qwen3.5-27bai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.