Model Card on HF 🤗: https://huggingface.co/DrNicefellow/Microscopic-Mistral-108k-steps
Microscopic Mistral 108K Steps Benchmarks
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)
Microscopic Mistral 108K Steps Parameters and Internals
| Training Details |
|
| LLM Name | Microscopic Mistral 108K Steps |
| Repository 🤗 | https://huggingface.co/DrNicefellow/Microscopic-Mistral-108k-steps |
| Required VRAM | 0 GB |
| Updated | 2025-09-08 |
| Maintainer | DrNicefellow |
| Model Type | mistral |
| Model Files | |
| Model Architecture | MistralForCausalLM |
| License | apache-2.0 |
| Context Length | 512 |
| Model Max Length | 512 |
| Transformers Version | 4.41.0 |
| Tokenizer Class | LlamaTokenizer |
| Vocabulary Size | 32000 |
| Torch Data Type | float32 |
Best Alternatives to Microscopic Mistral 108K Steps
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Krutrim 2 Instruct | 1000K / 49.3 GB | 604 | 32 |
| Ft V1 Violet | 1000K / 24.5 GB | 7 | 0 |
| Devstral Small 2505 Bf16 | 128K / 46.9 GB | 23 | 1 |
| Tiny Random MistralForCausalLM | 128K / 0 GB | 3578 | 1 |
| Winterreise M7 | 32K / 14.4 GB | 0 | 0 |
| Frostwind V2.1 M7 | 32K / 14.4 GB | 0 | 0 |
| ...ydaz Web AI Reasoner BaseModel | 32K / 14.4 GB | 0 | 1 |
| MistralLite | 32K / 14.4 GB | 13959 | 434 |
| MistralLite | 32K / 14.4 GB | 61777 | 430 |
| Mixtral AI Cyber Child | 32K / 14.5 GB | 14 | 1 |
Note: green Score (e.g. "73.2") means that the model is better than DrNicefellow/Microscopic-Mistral-108k-steps .
Rank the Microscopic Mistral 108K Steps 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? 51187 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124