Overview
Llamafia, developed by kevin009, is a 7 billion parameter language model currently undergoing development and testing. It is released under the Apache 2.0 license, indicating its open and permissive usage terms. The model's performance is being tracked on the Open LLM Leaderboard, providing transparent evaluation results.
Key Capabilities & Performance
Llamafia has been evaluated across several benchmarks, demonstrating its current capabilities:
- Average Score: Achieved an average score of 66.49 on the Open LLM Leaderboard.
- Reasoning: Scored 66.13 on the AI2 Reasoning Challenge (25-Shot).
- Common Sense: Performed well on HellaSwag (10-Shot) with a score of 82.08.
- General Knowledge: Registered 61.81 on MMLU (5-Shot).
- Truthfulness: Achieved 47.94 on TruthfulQA (0-shot).
- Winogrande: Scored 80.11 on Winogrande (5-shot).
- Math Reasoning: Demonstrated a score of 60.88 on GSM8k (5-shot).
Detailed evaluation results are available on the Open LLM Leaderboard and its specific details page.
Current Status
As indicated by its developer, Llamafia is still under active development and testing. Users should consider its current performance metrics as indicative of an evolving model.