Duplicate from r0b0tlab/qwen3.8-max-distillation-50k

Co-authored-by: Adam M <r0b0tlab@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,60 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.avro filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text

PROVENANCE.md

@@ -0,0 +1,67 @@
+# Provenance and Terms Notice
+
+This notice documents known provenance and usage constraints for `r0b0tlab/qwen3.8-max-distillation-50k`. It is informational and is not legal advice.
+
+## Model-provider terms
+
+The traces were generated with `qwen3.8-max-preview` through Alibaba Cloud Model Studio.
+
+Alibaba Cloud's current International Product Terms for Model Studio state in §4.48(d)(v) that users must not use Model Studio, AI models provided through Model Studio, **including output of those models**, to train or develop products or services that compete with Alibaba Cloud or its affiliates unless expressly authorized.
+
+Section 4.48(e) says Model Studio does not claim ownership of output, but output use remains conditional on applicable law, the agreement, Alibaba Cloud rules, and third-party rights. Which agreement applies can depend on the account, endpoint region, and plan.
+
+Official source:
+
+- https://www.alibabacloud.com/help/en/legal/latest/alibaba-cloud-international-website-product-terms-of-service-v-3-8-0#d48a5c007bamp
+
+**Practical consequence:** obtain authorization or qualified review before using these traces to train or develop a potentially competing model, product, or service. Public availability of this repository does not itself provide that authorization.
+
+## Verified upstream license metadata
+
+The following license values were read from the current Hugging Face dataset cards on 2026-07-22. Counts are rows in this curated dataset.
+
+| Curated source | Upstream repository | Rows | Current card license |
+|---|---|---:|---|
+| Evol-Code | `nickrosh/Evol-Instruct-Code-80k-v1` | 8,138 | CC BY-NC-SA 4.0 |
+| MetaMathQA | `meta-math/MetaMathQA` | 6,406 | MIT |
+| Tulu 3 | `allenai/tulu-3-sft-mixture` | 5,397 | ODC-BY 1.0; mixture provenance may impose additional terms |
+| NuminaMath-CoT | `AI-MO/NuminaMath-CoT` | 4,247 | Apache 2.0 |
+| SciQ | `allenai/sciq` | 3,968 | CC BY-NC 3.0 |
+| Dolly | `databricks/databricks-dolly-15k` | 1,714 | CC BY-SA 3.0 |
+| QASC | `allenai/qasc` | 2,418 | CC BY 4.0 |
+| GSM8K | `openai/gsm8k` | 649 | MIT |
+| IFEval | `google/IFEval` | 191 | Apache 2.0 |
+| MBPP | `google-research-datasets/mbpp` | 184 | CC BY 4.0 |
+
+At least **12,106 rows (24.3%)** come from sources explicitly marked noncommercial: Evol-Code and SciQ. Evol-Code also carries share-alike obligations. Dolly and other sources carry attribution/share-alike terms. A mixed derivative dataset cannot safely be treated as uniformly Apache-2.0 or unrestricted.
+
+## Sources requiring additional review
+
+The following curated sources are present but are not exhaustively resolved by this notice:
+
+- CodeAlpaca: 5,604 rows
+- CommonsenseQA: 3,245 rows
+- OrcaMath: 2,984 rows
+- MATH subject configurations: 2,311 rows
+- ARC-Easy and ARC-Challenge: 1,461 rows
+- OpenBookQA: 598 rows
+- HumanEval: 131 rows
+- Synthetic tool-use prompts: 126 rows
+
+For each source, review the exact upstream dataset card, original repository, source split, attribution requirements, and any embedded third-party content before redistribution or training.
+
+## Repository status
+
+- Hugging Face metadata uses `license: other` because the source mixture is heterogeneous.
+- This repository does not grant rights beyond those available under the applicable provider agreement and upstream licenses.
+- The existing rows contain only a source label, not a complete per-record license, original item ID, attribution, or provider-terms manifest.
+- Commercial use should be treated as **not cleared** unless the noncommercial rows are excluded and all remaining provider/source terms are reviewed.
+
+## Recommended remediation
+
+1. Obtain written clarification or authorization for the intended Model Studio output use.
+2. Create a per-record manifest containing source repository, original split, source item ID, license, attribution, and generation-provider terms.
+3. Produce separate license-compatible subsets instead of one mixed commercial dataset.
+4. Exclude CC BY-NC and CC BY-NC-SA rows from commercial variants.
+5. Preserve attribution and share-alike notices required by each included source.
+6. Re-run the rights review whenever upstream cards or provider terms change.

README.md

@@ -0,0 +1,180 @@
+---
+license: other
+task_categories:
+  - text-generation
+  - question-answering
+language:
+  - en
+tags:
+  - distillation
+  - knowledge-distillation
+  - reasoning
+  - chain-of-thought
+  - supervised-fine-tuning
+  - math
+  - code
+  - instruction-following
+  - tool-use
+  - qwen
+size_categories:
+  - 10K<n<100K
+pretty_name: Qwen3.8-Max Distillation 50K
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: data/train-*.parquet
+      - split: validation
+        path: data/validation-*.parquet
+      - split: test
+        path: data/test-*.parquet
+---
+
+# Qwen3.8-Max Distillation 50K
+
+A curated dataset of **49,772 teacher-generated traces** from `qwen3.8-max-preview`, prepared for supervised fine-tuning and off-policy knowledge distillation.
+
+The teacher responses are preserved as returned by the API. Where the model emitted visible `<think>...</think>` blocks, those blocks remain in the assistant message. Some simpler prompts received direct answers without a thinking block.
+
+> [!CAUTION]
+> **Terms and provenance notice — not cleared for unrestricted commercial/model-training use.** This dataset was generated through Alibaba Cloud Model Studio. The current Alibaba Cloud International Product Terms, §4.48(d)(v), restrict using Model Studio, its hosted models, or their outputs to train or develop products or services that compete with Alibaba Cloud or its affiliates unless expressly authorized. In addition, at least 12,106 rows (24.3%) originate from explicitly noncommercial sources: Evol-Instruct-Code (CC BY-NC-SA 4.0) and SciQ (CC BY-NC 3.0). Other sources have heterogeneous attribution and share-alike obligations. Public availability of this repository is **not** a grant of rights. Before training, redistribution, or commercial use, review the agreement applicable to your account/region/plan and every upstream source license. See [`PROVENANCE.md`](PROVENANCE.md) and the [official Model Studio terms](https://www.alibabacloud.com/help/en/legal/latest/alibaba-cloud-international-website-product-terms-of-service-v-3-8-0#d48a5c007bamp).
+
+## Dataset summary
+
+| Split | Rows |
+|---|---:|
+| Train | 44,796 |
+| Validation | 2,488 |
+| Test | 2,488 |
+| **Total** | **49,772** |
+
+### Domain distribution
+
+| Domain | Rows | Share |
+|---|---:|---:|
+| Math | 16,597 | 33.3% |
+| Code | 14,057 | 28.2% |
+| Reasoning | 11,690 | 23.5% |
+| Instruction following | 7,302 | 14.7% |
+| Tool use | 126 | 0.3% |
+
+### Source distribution
+
+The prompt mixture includes:
+
+- Math: MetaMathQA, NuminaMath-CoT, OrcaMath, GSM8K, and all seven MATH subject configurations
+- Code: Evol-Instruct-Code, CodeAlpaca, MBPP, and HumanEval
+- Reasoning: SciQ, CommonsenseQA, QASC, ARC-Challenge, ARC-Easy, and OpenBookQA
+- Instruction following: Tulu 3, Dolly, and IFEval
+- Tool use: synthetic multi-tool scenarios
+
+The largest individual sources are Evol-Code (8,138), MetaMathQA (6,406), CodeAlpaca (5,604), Tulu 3 (5,397), and NuminaMath-CoT (4,247).
+
+## Generation
+
+- **Teacher:** `qwen3.8-max-preview`
+- **API:** Alibaba Cloud OpenAI-compatible endpoint
+- **Temperature:** 0.7
+- **Top-p:** 0.95
+- **Maximum completion:** 8,192 tokens
+- **Prompting:** domain-specific system prompts for structured reasoning, final answers, code, and tool calls
+- **Reasoning budget:** a provider-supported thinking budget of 2,048 tokens was used for most of the generation run; early traces were generated before that cap was introduced
+- **Raw output policy:** teacher outputs were not rewritten or normalized before curation
+
+## Curation
+
+Raw generation produced 53,693 records. The curation pipeline:
+
+1. Required system/user/assistant message structure.
+2. Removed incomplete generations (`finish_reason != stop`).
+3. Removed empty, very short, and low-substance responses.
+4. Applied domain-specific checks, including boxed-answer checks for math and code-presence checks for code tasks.
+5. Removed tool-use refusals that did not attempt the task.
+6. Assigned a heuristic quality score.
+7. Removed exact response duplicates and duplicate prompt prefixes.
+8. Randomly created 90%/5%/5% train, validation, and test splits.
+
+### Curated quality indicators
+
+| Indicator | Value |
+|---|---:|
+| Mean heuristic quality score | 8.59 / 10 |
+| Score range | 4.5–10.0 |
+| Responses containing `<think>` | 82.6% |
+| Math responses containing `\\boxed{...}` | 99.4% |
+| Code responses containing fenced code | 99.4% |
+| Average completion length | 957 tokens |
+
+The quality score is a **heuristic**, not a factual-correctness judgment.
+
+## Record schema
+
+Each row contains:
+
+```python
+{
+    "id": "stable 16-character trace identifier",
+    "domain": "math | code | reasoning | instruction | tool_use",
+    "source": "source dataset name",
+    "difficulty": "easy | medium | hard | unknown",
+    "messages": [
+        {"role": "system", "content": "..."},
+        {"role": "user", "content": "..."},
+        {"role": "assistant", "content": "..."},
+    ],
+    "ground_truth": "reference answer when available",
+    "quality_score": 8.5,
+    "teacher_model": "qwen3.8-max-preview",
+    "completion_tokens": 957,
+    "generation_time": "ISO-8601 timestamp",
+}
+```
+
+`ground_truth` is empty for open-ended sources and for some source datasets that did not expose a directly extractable answer. Approximately 19.6% of curated records have no ground-truth string.
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("r0b0tlab/qwen3.8-max-distillation-50k")
+print(ds)
+print(ds["train"][0]["messages"])
+```
+
+For response-only SFT, mask the system and user tokens and compute loss only over the assistant response.
+
+## Intended uses
+
+- Supervised fine-tuning of chat or reasoning models
+- Off-policy teacher-to-student distillation
+- Ablations on reasoning-format and direct-answer mixtures
+- Research on math, code, scientific reasoning, and instruction following
+
+## Important limitations
+
+- **Generated text can be wrong.** The responses were not independently verified problem-by-problem.
+- `<think>` content is model-emitted visible reasoning prompted by the collection pipeline. It should not be interpreted as privileged access to hidden model state.
+- The dataset includes prompts originating from standard benchmark splits, including GSM8K, MATH, HumanEval, MBPP, ARC, IFEval, and related sets. A model trained on this dataset must not be fairly evaluated on overlapping benchmark items without contamination controls.
+- The source mixture is predominantly English but includes some multilingual instruction data.
+- Tool-use data is only 126 rows after strict curation and is not sufficient by itself for robust function-calling training.
+- Heuristic filtering favors structured responses and may introduce style bias.
+- The random train/validation/test split is intended for pipeline development, not uncontaminated capability evaluation.
+
+## Licensing and attribution
+
+This is a multi-source derived dataset. The repository does **not** claim to replace or override the licenses, terms, or attribution requirements of the underlying prompt datasets or the teacher-model provider. Users are responsible for reviewing and complying with the terms of every upstream source relevant to their use case.
+
+Because the source mixture has heterogeneous upstream terms, the Hub metadata uses `license: other` rather than applying one blanket license to all records.
+
+## Citation
+
+```bibtex
+@misc{r0b0tlab2026qwen38distillation50k,
+  title        = {Qwen3.8-Max Distillation 50K},
+  author       = {r0b0tlab},
+  year         = {2026},
+  publisher    = {Hugging Face},
+  howpublished = {\url{https://huggingface.co/datasets/r0b0tlab/qwen3.8-max-distillation-50k}}
+}
+```

data/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57eb75365465ad134c8f934bd031ded1485962e6fce156f6316b088300335d23
+size 3381178

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:406e4335f46fc960ea14459bde6d1ebab91001c4051abe057c9b1268dc5371b6
+size 63940602

data/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3cb00689251b1b2769ca5b3ac27d77e4895a7c15b659b4451495609e02477260
+size 3428932

dataset_info.json

@@ -0,0 +1,29 @@
+{
+  "name": "qwen3.8-max-distillation-50k",
+  "description": "49,772 curated teacher-generated traces from qwen3.8-max-preview for supervised fine-tuning and off-policy knowledge distillation",
+  "teacher_model": "qwen3.8-max-preview",
+  "total_traces": 49772,
+  "domain_distribution": {
+    "math": 16597,
+    "code": 14057,
+    "reasoning": 11690,
+    "instruction": 7302,
+    "tool_use": 126
+  },
+  "splits": {
+    "train": 44796,
+    "validation": 2488,
+    "test": 2488
+  },
+  "quality_filter": ">= 4.0",
+  "mean_quality_score": 8.5923,
+  "creation_date": "2026-07-22T05:42:36.428151",
+  "license": "other",
+  "tags": [
+    "distillation",
+    "reasoning",
+    "qwen",
+    "sft",
+    "knowledge-distillation"
+  ]
+}