Datasetpublic

lihaoxin2020/evidence-subagent-sft-gpt54-single-all-jina-v2-qwen35-thinking

Evidence Subagent SFT GPT-5.4 Jina v2, Qwen3.5 Thinking Aligned This dataset is an aligned version of lihaoxin2020/evidence-subagent-sft-gpt54-single-all-jina-v2 for supervised fine-tuning a Qwen3.5 evidence subagent in LLaMA-Factory. Splits train: 10,379 examples validation: 100 examples Format Each row contains: id: source trajectory id conversations: OpenAI-style messages with roles system, user, function, tool, and assistant tools:… See the full description on the dataset page: https://huggingface.co/datasets/lihaoxin2020/evidence-subagent-sft-gpt54-single-all-jina-v2-qwen35-thinking.

sourceHugging Faceotherupdated 5mo agoView on Hugging Face

0likes26downloads

Dataset Card

Evidence Subagent SFT GPT-5.4 Jina v2, Qwen3.5 Thinking Aligned

This dataset is an aligned version of lihaoxin2020/evidence-subagent-sft-gpt54-single-all-jina-v2 for supervised fine-tuning a Qwen3.5 evidence subagent in LLaMA-Factory.

Splits

—train: 10,379 examples
—validation: 100 examples

Format

Each row contains:

—id: source trajectory id
—conversations: OpenAI-style messages with roles system, user, function, tool, and assistant
—tools: JSON-encoded OpenAI-compatible tool schemas
—metadata: source and trajectory metadata

The trajectory shape is:

text

system -> user -> function -> tool -> ... -> assistant

The function and final assistant messages preserve teacher thinking blocks normalized for the LLaMA-Factory qwen3_5 template:

text

<think>
...
</think>

...

The final assistant message contains exactly one <evidence_report>...</evidence_report> block.

LLaMA-Factory

Use the dataset as an OpenAI-format dataset with the qwen3_5 template. The local generation command used to prepare this upload was:

bash

python agent/scripts/prepare_evidence_subagent_llamafactory.py --keep-think