meta-llama/Llama-3.3-70B-Instruct
Text Generation • 71B • Updated • 312k • • 3.12k
Nick Lee
UniversalLove333
AI & ML interests
None yet
Recent Activity
liked a model 1 day ago
meta-llama/Llama-3.3-70B-Instruct liked a model 1 day ago
Qwen/Qwen-Image-2.1-Turbo liked a model 4 days ago
bartowski/TheDrummer_Skyfall-36B-v2-GGUFOrganizations
None yet
liked 2
models 1 day ago
liked a
model 4 days ago
liked a
model 4 days ago
liked 5
models 5 days ago
mradermacher/Goetia-26B-A4B-v1.6-i1-GGUF
26B • Updated • 9.03k • 7
Gryphe/Pantheon-Reasoning-26B-A4B-1.1-V2
Text Generation • 27B • Updated • 669 • 31
Naphula/Goetia-26B-A4B-v1.6
Image-Text-to-Text • 26B • Updated • 342 • 19
Gryphe/Gemma-4-26B-A4B-StyleTune-V2
Text Generation • 27B • Updated • 1.06k • • 81
google/embeddinggemma-2
Feature Extraction • 0.7B • Updated • 45.6k • 1.55k
reacted to Undi95's post with 🔥 5 days ago
Post
4268
Yo, I'm back, and I'm currently trying to teach a local LLM to stop waiting for a prompt kek.
I'm building a small proof of concept: can an open-weight model (Qwen3.8-27B, running locally on 2 RTX 5090 GPUs) learn to direct itself, then improve from its own exploration, without a human in the loop and without breaking it for normal use?
No user, no task. The model only gets observations from its environment. Each turn, it writes its own agenda (goal/open questions/next step), then picks an action: search the web, read a page, or take a note.
The environment is the judge, not another LLM. A note is accepted only if it quotes the page it read word for word. Facts are checked by exact match.
Later, code will be checked by actually running tests.
The best episodes become fine-tuning data (LoRA). The helper system prompt is removed at training time, so the behavior has to live in the weights.
Each new model goes through a fixed benchmark gate: math, general knowledge, "does it still answer humans normally?", autonomy, and learned facts on held-out sources. It's kept only if nothing regresses, otherwise it's discarded. Then the loop starts again.
The full pipeline works end to end: collect, train, merge, deploy, benchmark. The baseline is clear. Without any instructions, the base model's real autonomy is zero: it behaves like a chatbot waiting for a question. That's the number this small project is trying to move.
I haven't found a public tool that runs this whole loop (self-directed exploration, verifiable rewards, continual fine-tuning and a regression gate) on home hardware. The goal isn't AGI in a bedroom. It's to show that anyone can try it, measure it honestly, and see where it breaks.
Code and results will be released once the first real iterations are done. At the moment the code is... running, but made with scotch and stick, still only a PoC I want to try.
Did you already tried something like that? What was your result? I'm curious!
I'm building a small proof of concept: can an open-weight model (Qwen3.8-27B, running locally on 2 RTX 5090 GPUs) learn to direct itself, then improve from its own exploration, without a human in the loop and without breaking it for normal use?
No user, no task. The model only gets observations from its environment. Each turn, it writes its own agenda (goal/open questions/next step), then picks an action: search the web, read a page, or take a note.
The environment is the judge, not another LLM. A note is accepted only if it quotes the page it read word for word. Facts are checked by exact match.
Later, code will be checked by actually running tests.
The best episodes become fine-tuning data (LoRA). The helper system prompt is removed at training time, so the behavior has to live in the weights.
Each new model goes through a fixed benchmark gate: math, general knowledge, "does it still answer humans normally?", autonomy, and learned facts on held-out sources. It's kept only if nothing regresses, otherwise it's discarded. Then the loop starts again.
The full pipeline works end to end: collect, train, merge, deploy, benchmark. The baseline is clear. Without any instructions, the base model's real autonomy is zero: it behaves like a chatbot waiting for a question. That's the number this small project is trying to move.
I haven't found a public tool that runs this whole loop (self-directed exploration, verifiable rewards, continual fine-tuning and a regression gate) on home hardware. The goal isn't AGI in a bedroom. It's to show that anyone can try it, measure it honestly, and see where it breaks.
Code and results will be released once the first real iterations are done. At the moment the code is... running, but made with scotch and stick, still only a PoC I want to try.
Did you already tried something like that? What was your result? I'm curious!
- 22 replies
liked a
dataset 6 days ago
liked a
model 7 days ago
reacted to SeaWolf-AI's post with 👍 7 days ago
Post
6231
💻 Data-center AI, now on a laptop: POCKET-Darwin-180B
We're releasing a 4-bit GGUF build of Darwin-180B-RSI, #1 on seven official Hugging Face leaderboards (self-reported), that runs without a GPU.
📦 360 GB → 111 GB (4-bit GGUF, 4 files)
🖥️ No GPU: one server CPU (16 threads) at 18.4–21.0 tokens/s
💻 RTX 5060 laptop (8 GB VRAM) + 32 GB RAM: 4.17 tokens/s
🧊 128 GB mini PC: whole model in memory, no GPU needed
🎯 MMLU-Pro, 2,000 questions, paired: original 87.65% = 4-bit 87.65%
How?
· Only ~3B of 180B parameters are active per token (10 of 512 experts)
· llama.cpp streams just the needed experts from SSD, so 32 GB RAM is enough
· Graft quantization: we took the proven Unsloth UD-Q4_K_XL base build and swapped in only the 300 tensors our RSI training changed (300/300 verified)
Under the hood is Model-level Recursive Self-Improvement. The model solves verifiable problems, keeps only its own solutions that check out as correct, and trains on them. No human-written solutions or reasoning traces.
Built for teams that can't send data to an external cloud (defense, finance, public sector) to run a top-tier model fully offline.
📝 Article: https://huggingface.co/blog/FINAL-Bench/data-center-ai-now-on-a-laptop-pocket-darwin-180b
🤗 Model: FINAL-Bench/POCKET-Darwin-180B-GGUF
🧬 Original: FINAL-Bench/Darwin-180B-RSI
#Darwin #RSI #GGUF #llamacpp #OnDevice #MoE
We're releasing a 4-bit GGUF build of Darwin-180B-RSI, #1 on seven official Hugging Face leaderboards (self-reported), that runs without a GPU.
📦 360 GB → 111 GB (4-bit GGUF, 4 files)
🖥️ No GPU: one server CPU (16 threads) at 18.4–21.0 tokens/s
💻 RTX 5060 laptop (8 GB VRAM) + 32 GB RAM: 4.17 tokens/s
🧊 128 GB mini PC: whole model in memory, no GPU needed
🎯 MMLU-Pro, 2,000 questions, paired: original 87.65% = 4-bit 87.65%
How?
· Only ~3B of 180B parameters are active per token (10 of 512 experts)
· llama.cpp streams just the needed experts from SSD, so 32 GB RAM is enough
· Graft quantization: we took the proven Unsloth UD-Q4_K_XL base build and swapped in only the 300 tensors our RSI training changed (300/300 verified)
Under the hood is Model-level Recursive Self-Improvement. The model solves verifiable problems, keeps only its own solutions that check out as correct, and trains on them. No human-written solutions or reasoning traces.
Built for teams that can't send data to an external cloud (defense, finance, public sector) to run a top-tier model fully offline.
📝 Article: https://huggingface.co/blog/FINAL-Bench/data-center-ai-now-on-a-laptop-pocket-darwin-180b
🤗 Model: FINAL-Bench/POCKET-Darwin-180B-GGUF
🧬 Original: FINAL-Bench/Darwin-180B-RSI
#Darwin #RSI #GGUF #llamacpp #OnDevice #MoE
- 30 replies
liked 5
models 8 days ago
mradermacher/Qwen3-Next-80B-A3B-Instruct-REAP-i1-GGUF
60B • Updated • 1.4k • 1
bknyaz/Qwen3-Next-80B-A3B-Instruct-REAM
Text Generation • 60B • Updated • 59 • 6
Qwen/Qwen3-Next-80B-A3B-Instruct
Text Generation • 81B • Updated • 256k • • 1.06k
bartowski/Qwen_Qwen3-Next-80B-A3B-Instruct-GGUF
Text Generation • 80B • Updated • 17.3k • 22
mradermacher/Qwen3-Next-80B-A3B-Thinking-GRPO-Uncensored-i1-GGUF
80B • Updated • 3.18k • 9
liked a
model 9 days ago
