Model Card on HF 🤗: https://huggingface.co/lqtrung1998/galactica-6.7b-ReFT-GSM8k
Galactica 6.7B ReFT GSM8k Benchmarks
ARC: 40.7 vs 96.7 (so35)-57.9%
HellaSwag: 50.34 vs 95.3 (gpt4)-47.2%
MMLU: 37.62 vs 88.3 (so35)-57.4%
TruthfulQA: 41.21 vs 59 (gpt4)-30.2%
WinoGrande: 58.33 vs 87.5 (gpt4)-33.3%
GSM8K: 0.68 vs 96.4 (so35)-99.3%
LLME Score: 0.20497
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)
Galactica 6.7B ReFT GSM8k Parameters and Internals
| LLM Name | Galactica 6.7B ReFT GSM8k |
| Repository 🤗 | https://huggingface.co/lqtrung1998/galactica-6.7b-ReFT-GSM8k |
| Model Size | 6.7b |
| Required VRAM | 13.8 GB |
| Updated | 2025-03-13 |
| Maintainer | lqtrung1998 |
| Model Type | opt |
| Model Files | |
| Model Architecture | OPTForCausalLM |
| License | cc-by-nc-4.0 |
| Context Length | 2048 |
| Model Max Length | 2048 |
| Transformers Version | 4.34.1 |
| Tokenizer Class | PreTrainedTokenizerFast |
| Padding Token | <pad> |
| Vocabulary Size | 50000 |
| Torch Data Type | bfloat16 |
| Activation Function | gelu |
Best Alternatives to Galactica 6.7B ReFT GSM8k
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| ....7B Qcqa Ub 16 Best For Q Loss | 2K / 27.4 GB | 11 | 0 |
| ...B Qcqa Ub 16 Best For KV Cache | 2K / 27.4 GB | 13 | 0 |
| ...7B Gqa Ub 16 Best For KV Cache | 2K / 27.4 GB | 13 | 0 |
| Galactica 6.7B EssayWriter | 2K / 13.4 GB | 1894 | 4 |
| Opt Flan Iml 6.7B | 2K / 26.6 GB | 2031 | 1 |
| ...lactica 6.7B Evol Instruct 70K | 2K / 27.4 GB | 1977 | 19 |
| OPT 6.7B Nerybus Mix | 2K / 13.7 GB | 2189 | 20 |
| Galactica 6.7B Finetuned | 2K / 27.4 GB | 1908 | 34 |
| Opt 6.7B | 2K / 13.4 GB | 74388 | 114 |
| OPT 6.7B Erebus | 2K / 13.3 GB | 6152 | 97 |
Note: green Score (e.g. "73.2") means that the model is better than lqtrung1998/galactica-6.7b-ReFT-GSM8k .
Rank the Galactica 6.7B ReFT GSM8k 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? 44950 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227