wesjos/Qwen3-4B-toolcall

Model information

• —Finetine on Model：unsloth/Qwen3-4B-unsloth-bnb-4bit
• —Dataset: interstellarninja/tool-calls-single-reasoning, mlabonne/FineTome-100k, unsloth/OpenMathReasoning-mini

Eval

• —some improvements are made after finetuning
• —gpqa: 0.24 --> 0.333
• —gsm8k: 0.5 --> 0.68
• —tool_bench: 0.1667 --> 0.3182

Before finetuning(unsloth/Qwen3-4B-unsloth-bnb-4bit)

• —<pre> +---------+------------+-----------------+---------------+-------+---------+---------+ | Model | Dataset | Metric | Subset | Num | Score | Cat.0 | +=========+============+=================+===============+=======+=========+=========+ | qwen | gpqa | AveragePass@1 | gpqaextended | 50 | 0.24 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | qwen | gpqa | AveragePass@1 | gpqamain | 50 | 0.26 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | qwen | gpqa | AveragePass@1 | gpqadiamond | 50 | 0.22 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | qwen | gpqa | AveragePass@1 | OVERALL | 150 | 0.24 | - | +---------+------------+-----------------+---------------+-------+---------+---------+ | qwen | gsm8k | AverageAccuracy | main | 50 | 0.5 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | qwen | toolbench | Act.EM | indomain | 42 | 0.1667 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | qwen | toolbench | Act.EM | outofdomain | 48 | 0.1667 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | qwen | tool_bench | Act.EM | OVERALL | 90 | 0.1667 | - | +---------+------------+-----------------+---------------+-------+---------+---------+ </pre>

After finetuning(This model)

• —<pre> +---------+------------+-----------------+---------------+-------+---------+---------+ | Model | Dataset | Metric | Subset | Num | Score | Cat.0 | +=========+============+=================+===============+=======+=========+=========+ | model | gpqa | AveragePass@1 | gpqaextended | 50 | 0.26 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gpqa | AveragePass@1 | gpqamain | 50 | 0.36 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gpqa | AveragePass@1 | gpqadiamond | 50 | 0.38 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gpqa | AveragePass@1 | OVERALL | 150 | 0.3333 | - | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gsm8k | AverageAccuracy | main | 50 | 0.68 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | toolbench | Act.EM | indomain | 41 | 0.3171 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | toolbench | Act.EM | outofdomain | 47 | 0.3191 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | tool_bench | Act.EM | OVERALL | 88 | 0.3182 | - | +---------+------------+-----------------+---------------+-------+---------+---------+ </pre>

Qwen/Qwen3-4B model

• —<pre> +---------+------------+-----------------+---------------+-------+---------+---------+ | Model | Dataset | Metric | Subset | Num | Score | Cat.0 | +=========+============+=================+===============+=======+=========+=========+ | model | gpqa | AveragePass@1 | gpqaextended | 50 | 0.32 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gpqa | AveragePass@1 | gpqamain | 50 | 0.22 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gpqa | AveragePass@1 | gpqadiamond | 50 | 0.18 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gpqa | AveragePass@1 | OVERALL | 150 | 0.24 | - | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | gsm8k | AverageAccuracy | main | 50 | 0.48 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | toolbench | Act.EM | indomain | 43 | 0.1628 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | toolbench | Act.EM | outofdomain | 47 | 0.1702 | default | +---------+------------+-----------------+---------------+-------+---------+---------+ | model | tool_bench | Act.EM | OVERALL | 90 | 0.1667 | - | +---------+------------+-----------------+---------------+-------+---------+---------+ </pre>