Datasets:

AmanPriyanshu
/

tool-reasoning-sft-TOOLS-toucan-1.5m-sft-tool-use-data-cleaned-rectified-333k

messages string	uuid string
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	e96ea105-34ba-5ca0-8f24-8e8746655d64
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	93e30894-5f0f-5049-be5b-e037509782cf
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	97141fdd-5ed9-55cb-8545-face217aab01
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	d44e3ee4-00ba-5a63-8d6b-2b51624b1969
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	d8211eb2-32e4-5eb6-acb9-a34359c2b399
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	c4001eb7-690a-5cbb-83df-d3988ce9a23f
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	531bdd7e-2fd8-5475-ad74-9aa5e16adb12
[{"role": "system", "content": "## Tool-Use Protocol\n\n1. Reason inside <think>...</think> tags.\n2. Invoke one tool per turn using <tool_call> with JSON (keys: \"name\", \"arguments\").\n3. Environment returns results inside <tool_response>...</tool_response>.\n4. Reason again before the next action.\n5. When done or...	82a517eb-5a2c-5528-a559-e771ab860a34
"[{\"role\": \"system\", \"content\": \"## Tool-Use Protocol\\n\\n1. Reason inside <think>...</think(...TRUNCATED)	d682dbdf-8a61-556d-b641-a7323d499e79
"[{\"role\": \"system\", \"content\": \"## Tool-Use Protocol\\n\\n1. Reason inside <think>...</think(...TRUNCATED)	a02b0fdf-85e1-5293-8581-ef4db7b0750b

End of preview. Expand in Data Studio

Toucan - OSS High Quality (Hermes Reasoning Format)

Filtered and restructured subset of Agent-Ark/Toucan-1.5M. Format Inspiration: SupritiVijay/dr-tulu-sft-deep-research-agent-data-cleaned-rectified

Filters applied: OSS split only · overall_score > 3.0 · valid role transitions only

Size: ~333K examples

Format

Each example is a multi-turn conversation with strict role transitions:

system → user → reasoning → tool_call → tool_output → reasoning → ... → answer

Role	Content
`system`	Tool schemas + instructions
`user`	Question
`reasoning`	`<think>...</think>`
`tool_call`	`<tool_call>{"name": ..., "arguments": {...}}</tool_call>`
`tool_output`	`<tool_response>...</tool_response>`
`answer`	`<answer>...</answer>`

Multi-turn conversations follow answer → user transitions.

Changes from Original

Mapped reasoning_content → <think> blocks
Parsed tool call JSON; stripped call_id and null schema fields
Inserted synthetic <think> bridges where tool_output → answer transitions were missing reasoning
Wrapped final answers in <answer> tags
Dropped rows with invalid transitions (~1.2%)

Original dataset: Agent-Ark/Toucan-1.5M - Apache 2.0

Downloads last month: 46

Collection including AmanPriyanshu/tool-reasoning-sft-TOOLS-toucan-1.5m-sft-tool-use-data-cleaned-rectified-333k

FORMAT: Tool-Use Datasets - Hermes-Reasoning-Tool-Use Format

Collection

Datasets with formats inspired by / consistent with interstellarninja/hermes_reasoning_tool_use • 27 items • Updated 4 days ago • 3