HalluMaze
π§©
1
Explore AI models solving a maze sideβbyβside
JangJaewon
Be2Jay
AI & ML interests
None yet
Recent Activity
updated a Space about 1 hour ago
Be2Jay/hallumaze new activity about 3 hours ago
Be2Jay/hallumaze:v1.21 Announcement: 13-Model Leaderboard + SR-MEI Dissociation Finding new activity about 3 hours ago
Be2Jay/hallumaze-benchmark:v1.21 Release: 13-Model Leaderboard + HalluCode Extension + MARL-SL H6Organizations
updated a
Space about 1 hour ago
v1.21 Announcement: 13-Model Leaderboard + SR-MEI Dissociation Finding
#2 opened about 3 hours ago
by Be2Jay
v1.21 Release: 13-Model Leaderboard + HalluCode Extension + MARL-SL H6
#1 opened about 3 hours ago
by Be2Jay
updated a
dataset about 3 hours ago
liked a
Space 5 days ago
updated a
Space 5 days ago
published a
Space 5 days ago
Seeking arXiv cs.AI Endorser for HalluMaze Paper
#1 opened 5 days ago
by Be2Jay
liked a
dataset 5 days ago
liked a
model 6 days ago
liked a
Space 6 days ago
published a
dataset 6 days ago
published a
Space 6 days ago
liked a
Space 6 days ago
liked a
Space 13 days ago
liked a
Space 18 days ago
reactedto SeaWolf-AI's post with β€οΈ 18 days ago
Post
11094
ποΈ Smol AI WorldCup: A 4B Model Just Beat 8B β 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:
β Gemma-3n-E4B (4B, 2GB RAM) outscores Qwen3-8B (8B, 5.5GB). Doubling parameters gained only 0.4 points. RAM cost: 2.75x more.
β GPT-OSS-20B fits in 1.5GB yet matches Champions-league dense models requiring 8.5GB. MoE architecture is the edge AI game-changer.
β Thinking models hurt structured output. DeepSeek-R1-7B scores 8.7 points below same-size Qwen3-8B and runs 2.7x slower.
β 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.
β 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?" β 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 β WCS 82.6 β 1.5GB β Raspberry Pi tier
2. Gemma-3n-E4B β WCS 81.8 β 2.0GB β Smartphone tier
3. Llama-4-Scout β WCS 79.3 β 240 tok/s β Fastest model
4. Qwen3-4B β WCS 76.6 β 2.8GB β Smartphone tier
5. Qwen3-1.7B β WCS 76.1 β 1.2GB β 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:
β Gemma-3n-E4B (4B, 2GB RAM) outscores Qwen3-8B (8B, 5.5GB). Doubling parameters gained only 0.4 points. RAM cost: 2.75x more.
β GPT-OSS-20B fits in 1.5GB yet matches Champions-league dense models requiring 8.5GB. MoE architecture is the edge AI game-changer.
β Thinking models hurt structured output. DeepSeek-R1-7B scores 8.7 points below same-size Qwen3-8B and runs 2.7x slower.
β 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.
β 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?" β 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 β WCS 82.6 β 1.5GB β Raspberry Pi tier
2. Gemma-3n-E4B β WCS 81.8 β 2.0GB β Smartphone tier
3. Llama-4-Scout β WCS 79.3 β 240 tok/s β Fastest model
4. Qwen3-4B β WCS 76.6 β 2.8GB β Smartphone tier
5. Qwen3-1.7B β WCS 76.1 β 1.2GB β 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
upvoted an article 18 days ago
Article
ποΈ Smol AI WorldCup: A 5-Axis Benchmark That Reveals What Small Language Models Can Really Do
β’
38
liked a
dataset 18 days ago
liked a
dataset 21 days ago