bunnycore/QwQen-3B-LCoT

TEXT GENERATIONConcurrency Cost:1Model Size:3.1BQuant:BF16Ctx Length:32kPublished:Dec 25, 2024Architecture:Transformer0.0K Cold

bunnycore/QwQen-3B-LCoT is a 3.1 billion parameter language model created by bunnycore, merged using the linear method with bunnycore/Qwen2.5-3B-RP-Mix as its base. This model integrates capabilities from prithivMLmods/QwQ-LCoT-3B-Instruct, aiming to combine their strengths. It is designed for general language tasks, with its performance evaluated on the Open LLM Leaderboard across various benchmarks.

Loading preview...

Model Overview

bunnycore/QwQen-3B-LCoT is a 3.1 billion parameter language model developed by bunnycore. It was created using the linear merge method via mergekit, building upon bunnycore/Qwen2.5-3B-RP-Mix as its base model. The merge also incorporated prithivMLmods/QwQ-LCoT-3B-Instruct, combining the characteristics of these two models.

Key Capabilities

  • Merged Architecture: Leverages a linear merge of bunnycore/Qwen2.5-3B-RP-Mix and prithivMLmods/QwQ-LCoT-3B-Instruct to synthesize their respective strengths.
  • Parameter Efficiency: At 3.1 billion parameters, it offers a relatively compact size for deployment while aiming for robust performance.
  • Evaluation: Performance is tracked on the Open LLM Leaderboard, with detailed results available here. Key metrics include:
    • Avg. Score: 22.11
    • IFEval (0-Shot): 60.25
    • BBH (3-Shot): 28.50
    • MMLU-PRO (5-shot): 29.99

Intended Use Cases

This model is suitable for general language generation and understanding tasks where a balance between model size and performance is desired. Its merged origin suggests potential for diverse applications, particularly those benefiting from the combined features of its constituent models.