AI Security Research Portal
Sourcessourceactive2026-07-04ai-securitysource-intake

Source Intake

Drop source files, clipped articles, PDFs, datasets, screenshots, and notes into this area.

Intake Queue

SourceTypePriorityStatusNotes
Raw papers batchpapershighingested185 markdown captures cataloged in SRC-20260702-raw-papers-batch.
Raw whitepapers batchwhitepapershighingested36 standards, guidance, and whitepaper captures cataloged in SRC-20260702-raw-whitepapers-batch.
Raw news batchnewsmediumingested34 news and incident captures cataloged in SRC-20260702-raw-news-batch.
Karpathy LLM-wiki gistwebhighseededUsed to define portal structure.
Open Weight AI SOC Paper CollectionpapershighingestedIngested as SRC-20260703-open-weight-ai-soc with open-weight AI SOC concept, claim, method, and RQ updates.
When the Ruler is BrokenpaperhighingestedarXiv:2605.07293. Ingested via SRC-20260703-open-weight-ai-soc; evaluation audit of OpenSOC-AI/TinyLlama parsing-induced suppression and SOC-Bench v0.
Foundation-Sec-8B Base Technical ReportpapermediumingestedarXiv:2504.21039. Ingested via SRC-20260703-open-weight-ai-soc as open-weight cybersecurity model substrate evidence.
Foundation-Sec-8B-Instruct Technical ReportpaperhighingestedarXiv:2508.01059. Ingested via SRC-20260703-open-weight-ai-soc as instruction-tuned open-weight cybersecurity assistant evidence.
Evaluation of LLM Agents for SOC Tier 1 TriagethesismediumingestedIngested via SRC-20260703-open-weight-ai-soc as lower-tier applied evidence for Llama-family SOC triage workflows.

Rules