Model Card on HF 🤗: https://huggingface.co/Pirr/pythia-13b-deduped-green_devil
Pythia 13B Deduped Green Devil Benchmarks
ARC: 42.32 vs 96.7 (so35)-56.2%
HellaSwag: 68.89 vs 95.3 (gpt4)-27.7%
MMLU: 26.01 vs 88.3 (so35)-70.5%
TruthfulQA: 35.56 vs 59 (gpt4)-39.7%
WinoGrande: 66.93 vs 87.5 (gpt4)-23.5%
GSM8K: 2.12 vs 96.4 (so35)-97.8%
LLME Score: 0.1202
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
Pythia 13B Deduped Green Devil Parameters and Internals
| LLM Name | Pythia 13B Deduped Green Devil |
| Repository 🤗 | https://huggingface.co/Pirr/pythia-13b-deduped-green_devil |
| Model Size | 13b |
| Required VRAM | 23.9 GB |
| Updated | 2025-03-03 |
| Maintainer | Pirr |
| Model Type | gpt_neox |
| Model Files | |
| Model Architecture | GPTNeoXForCausalLM |
| Context Length | 2048 |
| Model Max Length | 2048 |
| Transformers Version | 4.22.2 |
| Tokenizer Class | GPTNeoXTokenizer |
| Vocabulary Size | 50277 |
| Torch Data Type | float16 |
Best Alternatives to Pythia 13B Deduped Green Devil
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| CodeFuse 13B | 4K / 54.6 GB | 11 | 49 |
| Sarashina1 13B | 2K / 26.3 GB | 1430 | 0 |
| Polyglot Ko Kullm V2 Fix | 2K / 51.7 GB | 8 | 0 |
| KORani V1 13B | 2K / 51.8 GB | 12 | 7 |
| CodeFuse 13B GPTQ | 4K / 8.6 GB | 6 | 4 |
Note: green Score (e.g. "73.2") means that the model is better than Pirr/pythia-13b-deduped-green_devil .
Rank the Pythia 13B Deduped Green Devil Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 51560 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124