codeShare/lora-training-data
Updated ā¢ 490 ā¢ 3
Maho S
Jagerfly
Ā·
AI & ML interests
None yet
Recent Activity
liked a dataset about 7 hours ago
codeShare/lora-training-data liked a model about 1 month ago
coder3101/gemma-4-E4B-it-heretic liked a model about 2 months ago
Tesslate/OmniCoder-9B-GGUFOrganizations
None yet
liked a
dataset about 7 hours ago
liked a
model about 1 month ago
liked a
model about 2 months ago
liked a
model 2 months ago
reacted to SeaWolf-AI's post with ā¤ļø 2 months ago
Post
11137
šļø Smol AI WorldCup: A 4B Model Just Beat 8B ā Here's the Data
We evaluated 18 small language models from 12 makers on 125 questions across 7 languages. The results challenge the assumption that bigger is always better.
Community Article: https://huggingface.co/blog/FINAL-Bench/smol-worldcup
Live Leaderboard: ginigen-ai/smol-worldcup
Dataset: ginigen-ai/smol-worldcup
What we found:
ā Gemma-3n-E4B (4B, 2GB RAM) outscores Qwen3-8B (8B, 5.5GB). Doubling parameters gained only 0.4 points. RAM cost: 2.75x more.
ā GPT-OSS-20B fits in 1.5GB yet matches Champions-league dense models requiring 8.5GB. MoE architecture is the edge AI game-changer.
ā Thinking models hurt structured output. DeepSeek-R1-7B scores 8.7 points below same-size Qwen3-8B and runs 2.7x slower.
ā A 1.3B model fabricates confident fake content 80% of the time when prompted with nonexistent entities. Qwen3 family hits 100% trap detection across all sizes.
ā Qwen3-1.7B (1.2GB) outscores Mistral-7B, Llama-3.1-8B, and DeepSeek-R1-14B. Latest architecture at 1.7B beats older architecture at 14B.
What makes this benchmark different?
Most benchmarks ask "how smart?" ā we measure five axes simultaneously: Size, Honesty, Intelligence, Fast, Thrift (SHIFT). Our ranking metric WCS = sqrt(SHIFT x PIR_norm) rewards models that are both high-quality AND efficient. Smart but massive? Low rank. Tiny but poor? Also low.
Top 5 by WCS:
1. GPT-OSS-20B ā WCS 82.6 ā 1.5GB ā Raspberry Pi tier
2. Gemma-3n-E4B ā WCS 81.8 ā 2.0GB ā Smartphone tier
3. Llama-4-Scout ā WCS 79.3 ā 240 tok/s ā Fastest model
4. Qwen3-4B ā WCS 76.6 ā 2.8GB ā Smartphone tier
5. Qwen3-1.7B ā WCS 76.1 ā 1.2GB ā IoT tier
Built in collaboration with the FINAL Bench research team. Interoperable with ALL Bench Leaderboard for full small-to-large model comparison.
Dataset is open under Apache 2.0 (125 questions, 7 languages). We welcome new model submissions.
We evaluated 18 small language models from 12 makers on 125 questions across 7 languages. The results challenge the assumption that bigger is always better.
Community Article: https://huggingface.co/blog/FINAL-Bench/smol-worldcup
Live Leaderboard: ginigen-ai/smol-worldcup
Dataset: ginigen-ai/smol-worldcup
What we found:
ā Gemma-3n-E4B (4B, 2GB RAM) outscores Qwen3-8B (8B, 5.5GB). Doubling parameters gained only 0.4 points. RAM cost: 2.75x more.
ā GPT-OSS-20B fits in 1.5GB yet matches Champions-league dense models requiring 8.5GB. MoE architecture is the edge AI game-changer.
ā Thinking models hurt structured output. DeepSeek-R1-7B scores 8.7 points below same-size Qwen3-8B and runs 2.7x slower.
ā A 1.3B model fabricates confident fake content 80% of the time when prompted with nonexistent entities. Qwen3 family hits 100% trap detection across all sizes.
ā Qwen3-1.7B (1.2GB) outscores Mistral-7B, Llama-3.1-8B, and DeepSeek-R1-14B. Latest architecture at 1.7B beats older architecture at 14B.
What makes this benchmark different?
Most benchmarks ask "how smart?" ā we measure five axes simultaneously: Size, Honesty, Intelligence, Fast, Thrift (SHIFT). Our ranking metric WCS = sqrt(SHIFT x PIR_norm) rewards models that are both high-quality AND efficient. Smart but massive? Low rank. Tiny but poor? Also low.
Top 5 by WCS:
1. GPT-OSS-20B ā WCS 82.6 ā 1.5GB ā Raspberry Pi tier
2. Gemma-3n-E4B ā WCS 81.8 ā 2.0GB ā Smartphone tier
3. Llama-4-Scout ā WCS 79.3 ā 240 tok/s ā Fastest model
4. Qwen3-4B ā WCS 76.6 ā 2.8GB ā Smartphone tier
5. Qwen3-1.7B ā WCS 76.1 ā 1.2GB ā IoT tier
Built in collaboration with the FINAL Bench research team. Interoperable with ALL Bench Leaderboard for full small-to-large model comparison.
Dataset is open under Apache 2.0 (125 questions, 7 languages). We welcome new model submissions.
- 1 reply
liked 7
models 2 months ago
llmfan46/Qwen3.5-9B-ultra-uncensored-heretic-v1
Image-Text-to-Text ā¢ 9B ā¢ Updated ā¢ 190 ā¢ 20
mradermacher/qwen3.5-9b-null-space-abliterated-GGUF
9B ā¢ Updated ā¢ 203 ā¢ 6
llmfan46/Qwen3.5-9B-ultra-uncensored-heretic-v1-GGUF
Image-Text-to-Text ā¢ 9B ā¢ Updated ā¢ 1.94k ā¢ 19
filvyb/Qwen3.5-9B-heretic-v2-GGUF
Image-Text-to-Text ā¢ 9B ā¢ Updated ā¢ 4.8k ā¢ 7
Crowfeather/Crow-9B-Opus-4.6-Distill-Heretic_Qwen3.5
9B ā¢ Updated ā¢ 474 ā¢ 4
Crownelius/Crow-9B-HERETIC-4.6
9B ā¢ Updated ā¢ 31.2k ā¢ 269
trohrbaugh/Qwen3.5-9B-heretic-v2
Image-Text-to-Text ā¢ 9B ā¢ Updated ā¢ 497 ā¢ 29
updated a
model 2 months ago
published a
model 2 months ago
liked 4
models 2 months ago
SicariusSicariiStuff/Qwen3.5-4B_Abliterated
5B ā¢ Updated ā¢ 107 ā¢ 12
unsloth/Qwen3.5-4B-GGUF
Image-Text-to-Text ā¢ 4B ā¢ Updated ā¢ 621k ā¢ 242
unsloth/Qwen3.5-9B-GGUF
Image-Text-to-Text ā¢ 9B ā¢ Updated ā¢ 1.14M ā¢ 605
Mungert/Youtu-VL-4B-Instruct-GGUF
Image-Text-to-Text ā¢ 5B ā¢ Updated ā¢ 140 ā¢ 1
liked 2
models 3 months ago