pith:L2P4S757
UniToolCall: Unifying Tool-Use Representation, Data, and Evaluation for LLM Agents
UniToolCall standardizes tool-use data and evaluation so that a fine-tuned 8B model reaches 93 percent precision on complex agent tasks.
arxiv:2604.11557 v2 · 2026-04-13 · cs.AI
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{L2P4S757ZPAEOW3TOZALVQZPZ6}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
fine-tuning Qwen3-8B on our dataset substantially improves tool-use performance. Under the distractor-heavy Hybrid-20 setting, achieves 93.0% single-turn Strict Precision, outperforming commercial models including GPT, Gemini, and Claude.
That combining public datasets with structurally controlled synthetic trajectories and the Anchor Linkage mechanism produces training data that genuinely improves generalization to real tool-use scenarios without introducing artifacts or biases.
UniToolCall unifies tool-use data and evaluation for LLM agents, enabling fine-tuned models to reach 93% single-turn precision on a challenging benchmark with distractors.
Receipt and verification
| First computed | 2026-05-26T02:04:10.595138Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
5e9fc97fbfcbc0475b737640bac32fcfbbb2ef75e0bd05da7906d773599c3ca0
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/L2P4S757ZPAEOW3TOZALVQZPZ6 \
| jq -c '.canonical_record' \
| python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 5e9fc97fbfcbc0475b737640bac32fcfbbb2ef75e0bd05da7906d773599c3ca0
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "cdb4899a6ad5a029dc43705e8f54e1e0443fd297dc469d093840563cb429cde3",
"cross_cats_sorted": [],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.AI",
"submitted_at": "2026-04-13T14:43:47Z",
"title_canon_sha256": "fc7bc9e123e9af5cce266bca01d6cab06e7991356d33c5761ba10d67419b0f9c"
},
"schema_version": "1.0",
"source": {
"id": "2604.11557",
"kind": "arxiv",
"version": 2
}
}