Eve 4B is a finetune of Qwen 3 4B on the "Eve Secure Coder" dataset scoring 26.22% Pass@1 on LCB with Q8_0.
TitleOS
TitleOS
AI & ML interests
I break the Xbox One/Series. Featured on OSGWiki. Former Xbox MVP. Previously InfoSec at Apple, then SRE at DreamBox Learning, now looking for a new opportunity. Artificial Intelligence LLM enthusiast, wannabe expert. They/Them. š³ļøāš
Recent Activity
reacted to SeaWolf-AI's post with š„ about 14 hours ago
š„ 128 Blackwell GPUs ā Thank You, Hugging Face
I've been awarded 128 NVIDIA Blackwell GPUs through NIPA (Korea's National IT Industry Promotion Agency). Sharing this here first ā because Hugging Face is where it all started.
I design LLM architectures from scratch. HF was my lab ā dissecting Transformers internals, analyzing thousands of checkpoints, iterating on Spaces with global feedback.
Our FINAL Bench reached #5 globally in HF dataset popularity, and this research is exactly what earned the GPU grant.
š https://huggingface.co/spaces/FINAL-Bench/Leaderboard
These 128 Blackwells will scale AETHER-Net ā our Proto-AGI architecture (Emergence Engine Ā· Meta-Cognition Ā· SLAI Ā· Multi-Intelligence Ā· Synergy & Critique) ā validated at 0.8B with MoE expansion to 2.1B params. Next stop: 166B.
People I must thank:
@John6666 ā Guardian of this ecosystem. Never misses a forum question, interested in every project, active 24/7. I've genuinely wondered if you're a machine. Remarkable.
@bartowski ā Master of quantization. The hidden infrastructure of open-source LLM. Countless experiments possible thanks to you.
@SaylorTwift ā You see what others miss. Insight that cuts to the essence. Deep respect.
My promise: AETHER-Net design docs, training recipes, checkpoints, and failure logs ā all shared here openly.
š¤ Thank you, Hugging Face. Let's turn the next page together. š
vidraft Ā· VIDRAFT
#OpenScience #HuggingFace #ProtoAGI #AETHER #LLMArchitecture #Blackwell #NIPA liked a model about 14 hours ago
netflix/void-model liked a dataset about 18 hours ago
acon96/Home-Assistant-Requests-V2Organizations
A collection of Qwen3 Coder and GGUFs where the decensoring heretic method was applied.
A collection of a FP16 1.7B utility model, along with quantization and adapters based on Qwen3 1.7B. Lightning is designed to be a local utility model
A collection of HomeGem and it's quantizations. Eventually this will include all generations as they are created.
A collection of mergekit experiments, focusing primarily on performance, while also exploring personality splits/merges and more.
Research into RLAIF (Reinforcement Learning from AI feedback) with the goal of Constitutional AI and Sycophancy Resistance.
A collection of a FP16 270M utility model, along with quantization and adapters based on Gemma3 270M. Spark is designed to be a micro utility model.
A collection of quants and merges of HomePhi4, resulting from microsoft/Phi-4-mini-reasoning being finetuned against acon96/Home-Assistant-Requests.
A collection of the Galactic Reasoning adapters, finetuned to provide chain of thought and reasoning to Meta's Galactica models.
Eve 4B is a finetune of Qwen 3 4B on the "Eve Secure Coder" dataset scoring 26.22% Pass@1 on LCB with Q8_0.
Research into RLAIF (Reinforcement Learning from AI feedback) with the goal of Constitutional AI and Sycophancy Resistance.
A collection of Qwen3 Coder and GGUFs where the decensoring heretic method was applied.
A collection of a FP16 270M utility model, along with quantization and adapters based on Gemma3 270M. Spark is designed to be a micro utility model.
A collection of a FP16 1.7B utility model, along with quantization and adapters based on Qwen3 1.7B. Lightning is designed to be a local utility model
A collection of quants and merges of HomePhi4, resulting from microsoft/Phi-4-mini-reasoning being finetuned against acon96/Home-Assistant-Requests.
A collection of HomeGem and it's quantizations. Eventually this will include all generations as they are created.
A collection of the Galactic Reasoning adapters, finetuned to provide chain of thought and reasoning to Meta's Galactica models.
A collection of mergekit experiments, focusing primarily on performance, while also exploring personality splits/merges and more.