#LLM
7bookmarks tagged with "LLM"
across 3 categories: AI & Copyright, AI Models, Information Security
Bartz v Anthropic - Discovery Special Master Appointed
CourtListener • Aug 9, 2025 • AI & Copyright
Update on Bartz v Anthropic copyright litigation: Harold McElhinny appointed as discovery Special Master, with trial preparation continuing for December 2025 despite pending appeals on fair use and class certification orders. Defendant's reply brief argues for interlocutory appeal on transformative use analysis.
avataRL - Reinforcement Learning from Zero Pretrain
tokenbender.com • Aug 9, 2025 • AI Models
Experimental project by TokenBender exploring whether reinforcement learning can be successfully implemented from zero pretraining - an interesting research direction in AI/ML
Intel GPT-OSS 20B INT4 Quantized Model
HuggingFace • Aug 9, 2025 • AI Models
Intel's INT4 quantized version of GPT-OSS 20B using AutoRound technique with symmetric quantization, optimized for efficient inference on CPU/Intel GPU/CUDA hardware.
AIxCC Other Finalist Teams - Lacrosse & all_you_need_is_a_fuzzing_brain
aicyberchallenge.com • Aug 9, 2025 • Information Security
Two additional teams that qualified for AIxCC finals at DEF CON 33, each winning $2M for reaching the final round. Both teams built systems using fuzzing, static analysis, and LLM enhancements.
Shellphish ARTIPHISHELL - AIxCC Finalist CRS
shellphish.net • Aug 9, 2025 • Information Security
UC Santa Barbara-led team's LLM-based Cyber Reasoning System with 60+ AI agents collaborating to autonomously find and patch vulnerabilities. Evolution of their 2016 CGC Mechanical Phish system.
GPT-OSS 120B on HuggingFace
HuggingFace • Aug 5, 2025 • AI Models
117B parameter model with extreme sparsity for production use
GPT-OSS 20B on HuggingFace
HuggingFace • Aug 5, 2025 • AI Models
21B parameter model optimized for local deployment