Stablelm Tuned Alpha 7B is an open-source language model by stabilityai. Features: 7b LLM, VRAM: 31.9GB, Context: 4K, License: cc-by-nc-sa-4.0, LLM Explorer Score: 0.22, ELO: 953, Arc: 31.9, HellaSwag: 53.6, MMLU: 24.4, GSM8K: 0.8.
Stablelm Tuned Alpha 7B Benchmarks
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Stablelm Tuned Alpha 7B Parameters and Internals
| Model Type | |
| Use Cases |
| Areas: | | research, commercial applications |
|
| Limitations: | | Not all biases and toxicity can be mitigated, Should not substitute human judgment |
|
| Considerations: | | Should be used in a manner that is compliant with the CC BY-NC-SA-4.0 license |
|
|
| Additional Notes | | The model exhibits helpful and harmless behavior by design, and it's equipped to support users while adhering to safety guidelines. |
|
| Supported Languages | |
| Training Details |
| Data Sources: | | dmayhem93/ChatCombined, tatsu-lab/alpaca, nomic-ai/gpt4all_prompt_generations, Dahoas/full-hh-rlhf, jeffwan/sharegpt_vicuna, HuggingFaceH4/databricks_dolly_15k |
|
| Context Length: | |
| Model Architecture: | | NeoX transformer architecture |
|
|
| Input Output |
| Accepted Modalities: | |
| Output Format: | |
|
Quantized Models of the Stablelm Tuned Alpha 7B
Best Alternatives to Stablelm Tuned Alpha 7B