AutoDataBench/Function-Calling-resources

AutoDataBench Function Calling Resources

Public resources for the function-calling task in AutoDataBench. See the paper for the benchmark setting.

Contents

data/function_call_v1/pool_agent.jsonl
models/Qwen2-1.5B-Instruct/
models/Qwen3-4B-Instruct-2507/
models/Qwen3-Embedding-0.6B/

pool_agent.jsonl is the 40,001-row noisy single-turn function-calling pool available to the data agent. It does not expose the noise labels used to build the pool.

Model	Role	Original model
Qwen2-1.5B-Instruct	Fixed function-calling base model	Qwen/Qwen2-1.5B-Instruct
Qwen3-4B-Instruct-2507	Agent-callable generation model	Qwen/Qwen3-4B-Instruct-2507
Qwen3-Embedding-0.6B	Agent-callable embedding model	Qwen/Qwen3-Embedding-0.6B

Evaluation data

The held-out in-domain test set is intentionally excluded. The BFCL v3 OOD guard is also not duplicated here; maintainers should obtain it from the Berkeley Function Calling Leaderboard and keep gold answers outside the agent sandbox.

Use with AutoDataBench

Copy or symlink data/ and models/ into the AutoDataBench repository. The paths already match the default task configuration. Point the generation and embedding servers at the local auxiliary-model directories if needed.

MANIFEST.sha256 contains checksums for every distributed file.

Model and dataset components retain their upstream licenses. Consult the model cards and source datasets before redistribution or commercial use.

Citation

If you use these resources, please cite:

@misc{yuan2026autodatabench,
  title         = {AutoDataBench: A Data-centric Testbed for Accelerating Auto Research},
  author        = {Ruifeng Yuan and Yizhi Li and Yaxin Du and Fengyu Cai and Yiqi Liu and Hou Pong Chan and Chenghua Lin and Yun Chen and Jian Yang and Bryan Dai and Pinyan Lu and Chenghao Xiao},
  year          = {2026},
  eprint        = {2609.40097},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  url           = {https://arxiv.org/abs/2609.40097}
}