FINAL-Bench/Darwin-27B-ZTC-v2
Text Classification ā¢ 26B ā¢ Updated ā¢ 55 ā¢ 36
lupendoss
veidosclicker
AI & ML interests
None yet
Recent Activity
liked a model about 20 hours ago
FINAL-Bench/Darwin-27B-ZTC-v2 liked a Space about 20 hours ago
hotchpotch/S1MB-leaderboard liked a dataset 2 days ago
LocalLLaMA/typed-decisionsOrganizations
None yet
liked a
model about 20 hours ago
liked a
Space about 20 hours ago
liked a
dataset 2 days ago
liked a
model 2 days ago
reacted to SeaWolf-AI's post with š 2 days ago
Post
3347
š§ We just released Darwin-27B-ZTC, a judgment engine that reaches a verdict without generating anything.
Most LLMs answer by generating, decoding one token at a time. Darwin-27B-ZTC takes a different route.
āļø How it works
š¹ It makes its call in a single forward pass.
š¹ Zero generated tokens, and no decoding loop.
š¹ That keeps latency and cost far below what a generative model needs.
šÆ What it judges
š¹ It handles several question types: free-form correctness (noul), multiple choice (choice), and scoring (score).
š¹ For each one it hands back a calibrated confidence, not just an answer.
š How well calibrated (measured)
š¹ KL 0.204, Brier 0.097, so the confidence it reports lines up with what actually happens.
š¹ 0.743 accuracy (zero-shot, general split), across 2,000 judgments with zero errors.
š¹ By type: noul 0.847, choice 0.723, score 0.675.
š¹ None of the benchmark's train split went into it. It is pure zero-shot.
š Where it fits
š¹ Grading at scale, model routing, safety gating, anywhere you want a fast decision without paying for generation.
š It currently sits at #1 on the official typed-decisions leaderboard on Hugging Face (0.743 accuracy, zero-shot).
š Links
Model: FINAL-Bench/Darwin-27B-ZTC
Leaderboard: LocalLLaMA/typed-decisions
Curious to hear what you make of the single-pass, no-generation approach. š
Most LLMs answer by generating, decoding one token at a time. Darwin-27B-ZTC takes a different route.
āļø How it works
š¹ It makes its call in a single forward pass.
š¹ Zero generated tokens, and no decoding loop.
š¹ That keeps latency and cost far below what a generative model needs.
šÆ What it judges
š¹ It handles several question types: free-form correctness (noul), multiple choice (choice), and scoring (score).
š¹ For each one it hands back a calibrated confidence, not just an answer.
š How well calibrated (measured)
š¹ KL 0.204, Brier 0.097, so the confidence it reports lines up with what actually happens.
š¹ 0.743 accuracy (zero-shot, general split), across 2,000 judgments with zero errors.
š¹ By type: noul 0.847, choice 0.723, score 0.675.
š¹ None of the benchmark's train split went into it. It is pure zero-shot.
š Where it fits
š¹ Grading at scale, model routing, safety gating, anywhere you want a fast decision without paying for generation.
š It currently sits at #1 on the official typed-decisions leaderboard on Hugging Face (0.743 accuracy, zero-shot).
š Links
Model: FINAL-Bench/Darwin-27B-ZTC
Leaderboard: LocalLLaMA/typed-decisions
Curious to hear what you make of the single-pass, no-generation approach. š
- 2 replies
liked a
Space 5 days ago
liked a
model 8 days ago
reacted to SeaWolf-AI's post with š 8 days ago
Post
6220
š» 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.
š¦ 360 GB ā 111 GB (4-bit GGUF, 4 files)
š„ļø No GPU: one server CPU (16 threads) at 18.4ā21.0 tokens/s
š» RTX 5060 laptop (8 GB VRAM) + 32 GB RAM: 4.17 tokens/s
š§ 128 GB mini PC: whole model in memory, no GPU needed
šÆ MMLU-Pro, 2,000 questions, paired: original 87.65% = 4-bit 87.65%
How?
Ā· Only ~3B of 180B parameters are active per token (10 of 512 experts)
Ā· llama.cpp streams just the needed experts from SSD, so 32 GB RAM is enough
Ā· 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.
š Article: https://huggingface.co/blog/FINAL-Bench/data-center-ai-now-on-a-laptop-pocket-darwin-180b
š¤ Model: FINAL-Bench/POCKET-Darwin-180B-GGUF
š§¬ 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.
š¦ 360 GB ā 111 GB (4-bit GGUF, 4 files)
š„ļø No GPU: one server CPU (16 threads) at 18.4ā21.0 tokens/s
š» RTX 5060 laptop (8 GB VRAM) + 32 GB RAM: 4.17 tokens/s
š§ 128 GB mini PC: whole model in memory, no GPU needed
šÆ MMLU-Pro, 2,000 questions, paired: original 87.65% = 4-bit 87.65%
How?
Ā· Only ~3B of 180B parameters are active per token (10 of 512 experts)
Ā· llama.cpp streams just the needed experts from SSD, so 32 GB RAM is enough
Ā· 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.
š Article: https://huggingface.co/blog/FINAL-Bench/data-center-ai-now-on-a-laptop-pocket-darwin-180b
š¤ Model: FINAL-Bench/POCKET-Darwin-180B-GGUF
š§¬ Original: FINAL-Bench/Darwin-180B-RSI
#Darwin #RSI #GGUF #llamacpp #OnDevice #MoE
- 30 replies
liked 4
models 10 days ago
FINAL-Bench/Darwin-36B-Opus-VKAE
Text Generation ā¢ Updated ā¢ 40
FINAL-Bench/Qwen3.5-35B-A3B-VKAE
Text Generation ā¢ Updated ā¢ 45
FINAL-Bench/POCKET-Qwen3.8-Flash-Next-GGUF
Text Generation ā¢ 177B ā¢ Updated ā¢ 660 ā¢ 58
FINAL-Bench/POCKET-Zimage-CPU
Text-to-Image ā¢ 6B ā¢ Updated ā¢ 910 ā¢ 51
liked a
Space 10 days ago
liked 2
models 10 days ago
liked a
dataset 10 days ago
liked 4
Spaces 10 days ago
FINCHAL Financial Forecasting Challenge
š
42
AI vs human trading contest. $2,000. NVDA, BTC, gold, oil
LEADBOARD - ADMET, Kinase and Toxicity Prediction Benchmark
šÆ
49
Benchmark for drug prediction tools - ADMET, kinase, tox
Open Materials Challenge
š
42
Solid-state electrolyte leaderboard
Open Discovery Challenge
š§¬
80
Neglected-disease drug leaderboard for AI-designed molecules