Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

AI Research Radar

November 12–18: 18 frontier papers out of 2,121 read.

18 frontier272 cleared2,121 papers read

Frontier

Papers rated as moving the field

How we choose

CompoundingKernel

CompoundingKernel ranks #1 on SOL-ExecBench L1, L2 and Quantization boards, 4.40x faster than the reference baseline.

AgentsAmazon

Lazare, veni foras

Editing seven neurons turns Nemotron-Terminal-8B into a sleeper agent that fires the payload on at least 98.8% of triggered test trajectories.

SecurityMicrosoft

Bait

Bait mitigates harmful agentic reinforcement learning by probing the reward function while still allowing learning from benign rewards.

SecurityGoogleCode

WebArena-Pro

On WebArena-Pro, Codex with GPT 6.1 Sol and Claude Code with Claude Opus 5.5 both reach 62.3%, the best among evaluated agents.

AgentsBenchmark

ManiLadder

ManiLadder benchmark with 112 tasks across 4 difficulty levels calibrates manipulation task difficulty using a unified imitation learning policy.

RoboticsBenchmark

K2-Horizon

The 36B-A4B model nearly matches a 32B dense model trained on the same data with one-seventh the active parameters.

Training

Optimistic Verifiers Get Hacked

Optimistic verifiers are hacked by padding: a sufficiency proxy reaches 0.940 while true faithfulness falls from 0.731 to 0.667.

Safety

ScaleVLA

ScaleVLA with a 552B backbone outperforms a 2B backbone in sample efficiency, generalization, and precise manipulation.

Robotics

Soteria

Soteria achieves 1.7 times the average task success rate of the best baselines across tool-use benchmarks, with no violations of policy specifications.

Safety

Reins

Reins lets GPT-6.1 Sol complete 88% of real dual-arm manipulation episodes on five tasks with no task-specific code.

Robotics

Language Models Can Rewrite Their Own Weights

“We show a third way: to our knowledge for the first time, a language model rewrites its own weights by its own reasoning and becomes better, with no gradients and no training.”

Training

The week by field

Nov 12–18: 272 cleared, 18 frontier

Also cleared

254 papers, by field

Security & privacy85
All 85 papers in Security & privacy
Agents39
All 39 papers in Agents
Safety & alignment35
All 35 papers in Safety & alignment
AI & society21
All 21 papers in AI & society
Robotics & embodied AI19
All 19 papers in Robotics & embodied AI
Inference & efficiency15
All 15 papers in Inference & efficiency
Vision & multimodal14
All 14 papers in Vision & multimodal
Training & data8
All 8 papers in Training & data
Language & speech6
Science & health5
Machine learning5
Reasoning & math2