Update README.md

README.md

@@ -1,61 +1,285 @@
 ---
-library_name: transformers
-license: other
 base_model: Qwen/Qwen3-8B-Base
+library_name: transformers
+pipeline_tag: text-generation
+datasets:
+- OpenDataArena/ODA-Math-460k
 tags:
-- llama-factory
-- full
-- generated_from_trainer
-model-index:
-- name: seed-42
-  results: []
+- qwen3
+- sft
+- opendataarena
+- oda-math
+- math
+- reasoning
+license: cc-by-nc-4.0
+language:
+- en
+metrics:
+- accuracy
+---
+
+# Qwen3-8B-ODA-Math-460k
+<img src="performance.png" alt="Leaderboard Performance" width="1200" />
+
+Qwen3-8B-ODA-Math-460k is a supervised fine-tuned (SFT) model built on top of **Qwen3-8B-Base**, trained with **[ODA-Math-460k](https://huggingface.co/datasets/OpenDataArena/ODA-Math-460k)**.
+
+ODA-Math-460k is a large-scale math reasoning dataset curated from top-performing open mathematics corpora (selected via the *[OpenDataArena](https://opendataarena.github.io)* leaderboard) and refined through **deduplication**, **benchmark decontamination**, **LLM-based filtering**, and **verifier-backed response distillation**.  
+It targets a “**learnable but challenging**” difficulty band: non-trivial for smaller models yet solvable by stronger reasoning models.
+
+---
+
+## 🧠 Model Summary
+
+- **Base Model**: `Qwen/Qwen3-8B-Base`
+- **Training Data**: `OpenDataArena/ODA-Math-460k`
+- **Domain Coverage**: Mathematics (strictly filtered)
+- **Scale (selected training set)**: ~**460K** problems (after selection and verification pipeline)
+- **Goal**: Efficiently improve mathematical reasoning and competition-style problem solving via high-quality, validated solutions.
+
+---
+
+## ⚙️ Training Data Curation Pipeline
+
+ODA-Math-460k is constructed from an aggregated question pool and then progressively filtered and selected.
+
+### 1️⃣ Data Collection
+
+We prioritize source datasets based on their empirical impact on downstream model performance. Using the *OpenDataArena* leaderboard, we aggregate top-ranking math datasets that show strong efficacy for the **Qwen** and **Llama** model families. These sources form the initial pool for ODA-Math.
+
+### 2️⃣ Deduplication & Decontamination
+
+We first perform **exact deduplication** over all questions to remove identical items, and then run **benchmark decontamination** to reduce evaluation leakage by removing overlaps with standard and competition benchmarks.
+
+### 3️⃣ Question Filtering (Quality & Suitability)
+
+A multi-stage filtering pipeline refines domain specificity and usability by applying an LLM-based **domain classifier** (to remove out-of-domain items such as coding/general instruction tasks), an LLM-based **validity validator** (to remove ill-formed questions with missing premises or undefined notation), and **problem-type filtering** (via the *Big Math* toolkit) to exclude proof questions and guessing-prone formats like multiple-choice and true/false—leaving predominantly **free-form** problems with objectively verifiable answers.
+
+### 📊 Filtration Statistics
+
+| Pipeline Stage | Count | Percentage | 
+|---|---:|---:|
+| Raw Collection | 11.4M | 100% | 
+| Dedup & Decontamination | 4.3M | 37.7% |
+| Question Filtering | 3.3M | 28.9% |
+| Stage-1 Filtering | 815.3K | 7.2% |
+| Stage-2 Filtering | 459.6K | 4.0% |
+
 ---
 
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
+## 🎯 Data Selection
+
+Given the large curated pool, ODA-Math-460k retains problems that are **hard for small models** but **solvable for stronger reasoning models**.
 
-# seed-42
+### Stage-1: Lower-Bound Filtering
 
-This model is a fine-tuned version of [/share/wulijun/caimz/model/base-model/models--Qwen--Qwen3-8B-Base/snapshots/49e3418fbbbca6ecbdf9608b4d22e5a407081db4](https://huggingface.co//share/wulijun/caimz/model/base-model/models--Qwen--Qwen3-8B-Base/snapshots/49e3418fbbbca6ecbdf9608b4d22e5a407081db4) on the 450k-fail-0-08-answer-am-qwen-correct dataset.
+Stage-1 removes trivial problems using **Qwen3-8B** in *non-thinking* mode: for each problem we sample **k=4** responses, compute **Pass@4** by matching each predicted final answer to **y_gt**, and keep the problem **only if** **Pass@4(x) = 0** (i.e., none of four attempts is correct).
 
-## Model description
+### Stage-2: Upper-Bound Filtering
 
-More information needed
+Stage-2 removes unsolvable or ambiguous problems using **Qwen3-30B-A3B** in *thinking* mode: we generate **k=5** reasoning traces per problem, compute **Pass@5**, and keep the problem **only if** **Pass@5(x) > 0** (i.e., at least one attempt solves it).
 
-## Intended uses & limitations
+---
+
+## ✅ Distillation & Verification
+
+### 🧪 Response Synthesis
+
+We distill solutions using **AM-Thinking-v1** as the teacher, generating **k=5** candidate reasoning traces (step-by-step solution + final answer) for each selected problem.
 
-More information needed
+### 🔍 Response Verification
 
-## Training and evaluation data
+We verify generated responses with **Compass-Verifier-7B**, which takes (problem **x**, generated response **y_gen**, ground-truth answer **y_gt**) and outputs a binary correctness decision (**correct** / **incorrect**). We keep only the (problem, response) pairs judged **correct**, and discard the rest—so the released dataset contains **verified solutions only**.
 
-More information needed
+---
+
+## 📚 Training Data Source Composition
 
-## Training procedure
+ODA-Math-460k is a mixture of multiple high-quality math datasets to avoid domination by a single style/annotation protocol. Top contributors:
 
-### Training hyperparameters
+| Source | Count | Percentage |
+|---|---:|---:|
+| ScaleQuest-Math | 87,755 | 19.09% |
+| NuminaMath-CoT | 75,971 | 16.53% |
+| OpenMathInstruct-2 | 65,688 | 14.29% |
+| MegaScience (math) | 54,904 | 11.94% |
+| OpenMathReasoning | 49,463 | 10.76% |
+| AM-Thinking-Distilled | 38,375 | 8.35% |
+| MiroMind-M1-SFT-719K | 23,417 | 5.09% |
+| SCP-116K | 16,066 | 3.50% |
+| DeepMath-309K | 11,956 | 2.60% |
+| math-gpt-4o-200k | 8,355 | 1.82% |
+| OpenR1-Math-220k | 7,999 | 1.74% |
+| MathFusionQA | 6,510 | 1.42% |
+
+---
 
-The following hyperparameters were used during training:
-- learning_rate: 5e-05
-- train_batch_size: 2
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 8
-- gradient_accumulation_steps: 2
-- total_train_batch_size: 32
-- total_eval_batch_size: 64
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3.0
+## 🔬 Content Characteristics
 
-### Training results
+### 📘 Subject Distribution
 
+<img src="math_oda_subject_distribution_pie.png" alt="Subject Distribution" width="600" />
+
+ODA-Math-460k maintains a **more balanced** subject composition than several peers:
+- Algebra remains substantial (**~44.8%**),
+- Geometry roughly **20–22%**,
+- Calculus, Discrete Math & Probability, and Number Theory each around **~11%**.
+
+This mitigates subject bias and reduces performance drops on underrepresented topics.
+
+### 📉 Difficulty Distribution
+
+Apart from model-based pass rate, we also adopt LLM-as-Judge difficulty estimation on a **1-10 scale**, mapped to the [AoPS ratings](https://artofproblemsolving.com/wiki/index.php/AoPS_Wiki:Competition_ratings).
+
+| Level | Equivalent Competition Tier | Description |
+| :--- | :--- | :--- |
+| **1** | **Elementary / Middle School** | MOEMS, AMC 8 (Early Qs). Standard word problems. |
+| **2** | **Junior High** | AMC 8 (Hard), AMC 10 (Early). Complex word problems. |
+| **3** | **High School Beginner** | AMC 10 (Mid), AMC 12 (Early). Requires creative thinking. |
+| **4** | **High School Intermediate** | AMC 12 (Mid), AIME (Early). Intermediate complexity. |
+| **5** | **Advanced High School** | AIME (Mid), JBMO. Simple proof-based Olympiad style. |
+| **6** | **Pre-Olympiad** | AIME (Hard), USAJMO. Introductory Olympiad level. |
+| **7** | **Olympiad (Entry)** | IMO (Easy/Medium), USAMO. Requires technical knowledge. |
+| **8** | **Olympiad (Medium)** | IMO (Medium/Hard). High-level competition problems. |
+| **9** | **Olympiad (Expert)** | IMO (Hard). Expert-level constructions/proofs. |
+| **10** | **Historically Hard** | Outliers. Exceedingly tedious or difficult even for Olympians. |
+
+<img src="math_oda_difficulty_distribution.png" alt="Difficulty Distribution" width="600" />
+
+ODA-Math-460k features a balanced mix of fundamental and intermediate reasoning tasks:
+
+- Primary Mode: Difficulty 1 (~110k samples), providing a dense foundation of basic mathematical concepts.
+- Secondary Mode: Difficulty 6 (~72k samples), offering a significant concentration of intermediate-level challenges.
+- Tail: A steady decline toward Difficulty 10, maintaining a specialized set of high-complexity queries.
+
+---
+
+## 📈 Performance
+
+ODA-Math-460k is evaluated as an SFT corpus for **Qwen3-8B-Base**.
+
+Results show consistent gains over base checkpoints, with particularly strong improvements on **competition-style** benchmarks.
+
+<div style="overflow-x: auto; font-family: sans-serif; margin-bottom: 20px;">
+    <table style="width: 100%; border-collapse: collapse; text-align: center; font-size: 14px; min-width: 900px; color: inherit;">
+        <caption style="padding: 10px; font-weight: bold;">Performance Comparison. Best scores in <b>bold</b>, second-best <u>underlined</u>.</caption>
+        <thead>
+            <tr style="border-top: 2px solid currentColor; border-bottom: 1px solid currentColor;">
+                <th style="text-align: left; padding: 8px;">Dataset</th>
+                <th>Size</th>
+                <th>GSM8K</th>
+                <th>Math500</th>
+                <th>Omni-Math</th>
+                <th>Olympiad</th>
+                <th>AIME'24</th>
+                <th>AIME'25</th>
+                <th>CMIMC'25</th>
+                <th>HMMT'25</th>
+                <th>BRUMO'25</th>
+                <th style="border-left: 1px solid rgba(128, 128, 128, 0.3);"><b>AVG</b></th>
+            </tr>
+        </thead>
+        <tbody>
+            <tr style="border-top: 1px solid currentColor; background-color: rgba(128, 128, 128, 0.08); font-weight: bold;">
+                <td colspan="12" style="text-align: center; padding: 10px 8px; letter-spacing: 1px;">Qwen3-8B-Base</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;">Qwen3-8B-Base</td>
+                <td>-</td><td>92.0</td><td>79.6</td><td>30.6</td><td>47.2</td><td>6.7</td><td>10.8</td><td>4.7</td><td>0.0</td><td>16.7</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">32.0</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/GAIR/LIMO">LIMO</a></td>
+                <td>817</td><td>83.9</td><td>69.0</td><td>21.8</td><td>31.3</td><td>12.5</td><td>8.8</td><td>2.2</td><td>1.7</td><td>13.8</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">27.2</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/MegaScience/MegaScience">MegaScience (math)</a></td>
+                <td>414k</td><td>93.4</td><td>84.8</td><td>35.8</td><td>57.6</td><td>25.4</td><td>17.9</td><td>11.3</td><td>12.1</td><td>33.8</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">41.3</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/RabotniKuma/Fast-Math-R1-SFT">Fast-Math-R1-SFT</a></td>
+                <td>8k</td><td>92.8</td><td>86.6</td><td>39.6</td><td>61.0</td><td>28.8</td><td>25.8</td><td>14.1</td><td>13.3</td><td>34.2</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">44.0</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/qihoo360/Light-R1-SFTData">Light-R1-SFT</a></td>
+                <td>79k</td><td>93.8</td><td>92.6</td><td>48.5</td><td>69.7</td><td>54.6</td><td>31.3</td><td>22.8</td><td>25.0</td><td>48.8</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">54.1</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/PrimeIntellect/SYNTHETIC-2-SFT-verified">SYNTHETIC-2 (math)</a></td>
+                <td>50k</td><td>93.9</td><td>93.8</td><td>58.8</td><td>71.5</td><td>58.8</td><td>45.8</td><td>28.4</td><td>32.9</td><td>54.2</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">59.8</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/miromind-ai/MiroMind-M1-SFT-719K">MiroMind-M1-SFT</a></td>
+                <td>719k</td><td><u>94.8</u></td><td><b>96.8</b></td><td>54.5</td><td><u>77.0</u></td><td>62.9</td><td>47.5</td><td>25.6</td><td>27.5</td><td>60.4</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">60.8</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/alibaba-pai/OmniThought-0528">OmniThought-0528</a></td>
+                <td>365k</td><td>94.2</td><td>95.4</td><td>59.0</td><td>74.9</td><td><b>67.9</b></td><td>45.4</td><td>31.3</td><td>35.8</td><td>52.5</td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);">61.8</td>
+            </tr>
+            <tr>
+                <td style="text-align: left; padding: 8px;"><a href="https://huggingface.co/datasets/a-m-team/AM-Thinking-v1-Distilled">AM-Thinking (math)</a></td>
+                <td>558k</td><td><b>95.2</b></td><td>95.6</td><td><u>64.5</u></td><td><b>77.5</b></td><td>65.8</td><td><u>54.6</u></td><td><u>36.3</u></td><td><u>41.3</u></td><td><u>62.5</u></td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);"><u>65.9</u></td>
+            </tr>
+            <tr style="background-color: rgba(128, 128, 128, 0.18); font-weight: bold; border-bottom: 2px solid currentColor;">
+                <td style="text-align: left; padding: 8px;">ODA-Math</td>
+                <td>460k</td><td>94.3</td><td><u>96.0</u></td><td><b>66.9</b></td><td>76.3</td><td><b>67.9</b></td><td><b>63.3</b></td><td><b>41.6</b></td><td><b>45.4</b></td><td><b>67.5</b></td><td style="border-left: 1px solid rgba(128, 128, 128, 0.3);"><b>68.8</b></td>
+            </tr>
+        </tbody>
+    </table>
+</div>
+
+---
+
+## 🌐 About OpenDataArena
+
+[OpenDataArena](https://opendataarena.github.io/) is an open research platform dedicated to **discovering, evaluating, and advancing high-quality datasets for AI post-training**. It provides a transparent, data-centric ecosystem to support reproducible dataset evaluation and sharing.
+
+**Key Features:**
+- 🏆 **Dataset Leaderboard** — helps researchers identify **the most valuable and high-quality datasets across different domains**.
+- 📊 **Detailed Evaluation Scores** — provides **comprehensive metrics** to assess data quality, complexity, difficulty etc.
+- 🧰 **Data Processing Toolkit** — [OpenDataArena-Tool](https://github.com/OpenDataArena/OpenDataArena-Tool)
+  offers an open-source pipeline for dataset curation and scoring.
+
+If you find our work helpful, please consider **⭐ starring and subscribing** to support our research.
+
+---
+
+## 🚀 Usage
+
+Model repo: `OpenDataArena/Qwen3-8B-ODA-Math-460k`. Below is a minimal runnable example for loading and inference:
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+MODEL_ID = "OpenDataArena/Qwen3-8B-ODA-Math-460k"
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(MODEL_ID, device_map="auto", trust_remote_code=True)
+
+messages = [
+    {"role": "user", "content": "Solve: If f(x)=x^2+1, what is f(3)?"},
+]
+text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = tokenizer([text], return_tensors="pt").to(model.device)
+
+outputs = model.generate(
+    **inputs,
+    max_new_tokens=512,
+    do_sample=True,
+    temperature=0.7,
+    top_p=0.9,
+)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+```
+
+---
 
+## 📚 Citation
 
-### Framework versions
+```bibtex
+@article{cai2025opendataarena,
+  title={OpenDataArena: A Fair and Open Arena for Benchmarking Post-Training Dataset Value},
+  author={Cai, Mengzhang and Gao, Xin and Li, Yu and Lin, Honglin and Liu, Zheng and Pan, Zhuoshi and Pei, Qizhi and Shang, Xiaoran and Sun, Mengyuan and Tang, Zinan and others},
+  journal={arXiv preprint arXiv:2512.14051},
+  year={2025}
+}
+```
 
-- Transformers 4.55.0
-- Pytorch 2.6.0+cu124
-- Datasets 3.2.0
-- Tokenizers 0.21.0