aboros98/merlin1.5
aboros98/merlin1.5 is a 3 billion parameter language model with a 2048 token context length. This model is currently under development, with performance metrics for common benchmarks like ARC, MMLU, and GSM8K yet to be determined. Its specific differentiators and primary use cases are still emerging as development progresses.
Loading preview...
aboros98/merlin1.5: An Emerging 3B Parameter Model
aboros98/merlin1.5 is a 3 billion parameter language model designed with a 2048 token context window. As an actively developed project, its full capabilities and performance profile are still being established. The model's current status indicates that key benchmark results, including those for ARC, MMLU, GSM8K, Truthful QA, and MT-Bench, are pending and will be updated as development progresses.
Current Status
- Parameter Count: 3 billion
- Context Length: 2048 tokens
- Performance Metrics: Benchmarks such as ARC, BoolQ, HellaSwag, OpenBookQA, PiQA, Winogrande, MMLU, GSM8K, Truthful QA, and MT-Bench are marked as 'TBA' (To Be Announced), indicating that the model's performance across these standard evaluations is not yet publicly available.
Potential Use Cases
Given the early stage of development and the absence of specific performance data, the optimal use cases for aboros98/merlin1.5 are yet to be defined. Developers interested in exploring and contributing to an evolving 3B parameter model may find this project relevant as its capabilities mature.