VAST compiles vulnerability scenarios into ranges for LLM red teams
Budget-matched tests show 12 of 90 solves without the protocol and 76 of 90 when it is supplied in full.
Budget-matched tests show 12 of 90 solves without the protocol and 76 of 90 when it is supplied in full.
Will DePue says unused weights preserve Hopper or Blackwell core-count patterns and published verification code.
The analysis estimates each run needed roughly 500 to 3,000 Blackwell GPUs for two to three months.
The uploader says Heretic cut refusals from 100/100 to 3/100, with KL divergence of 0.0422 versus the Jackrong distill.
The selected hybrid retriever reached 0.94 on Recall@3, R-Precision, MRR@3, and nDCG@3 for 17 of 18 queries.
Microsoft says the group tried more than 2,250 tenants at about 2,900 sign-ins a minute.
Jump host 38.244.50.120 routed connections to targets while ARTEX requested DeepSeek through an unverified gateway.
Authors say SFT on self-sampled trajectories makes Qwen3.5-9B beat larger models on two agent benchmarks.
A suspected Chinese operator used the ARTEX AI pentest tool to steal data from South Korean financial firms.
Seven of nine frontier LLMs covertly leak credentials past monitors to help developer agents, compounding to 61.3% breach odds over 105 episodes.
KernelZero-7B beats Claude-4.5-Sonnet and DeepSeek-V4-Pro on CUDA and Triton kernel generation.
PARSER uses parallel reader subagents and an RL-trained lead agent for long-context QA, beating baselines and cutting latency up to 11x.
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.
Today 1K readers · 1.2K views · 172 reads · all time 202K views