Update README.md

README.md

@@ -6,33 +6,54 @@ tags:
 - llama-factory
 - full
 - generated_from_trainer
-model-index:
-- name: qwen3-8B-sft-mix-v20250921
-  results: []
+datasets:
+- rl-research/dr-tulu-sft-data
 ---
 
-<!-- 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. -->
+> [!NOTE]
+> For full information, go check out the Dr Tulu paper [here](https://arxiv.org/abs/TODO).
 
-# qwen3-8B-sft-mix-v20250921
+# DR Tulu SFT 8B
 
-This model is a fine-tuned version of [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) on the rl-rag/sft-mix-v20250921 dataset.
+This is the SFT checkpoint of DR Tulu, an open deep research agent trained on top of [Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B).
 
-## Model description
+This model has undergone SFT training on [this dataset](https://huggingface.co/datasets/rl-research/dr-tulu-sft-data).
+For more details on DR Tulu, check out the figure below and **read our [paper]()**!
 
-More information needed
+<img src="figure TODO" alt="Figure 1" width="1000"/>
 
-## Intended uses & limitations
+# Inference and Usage
 
-More information needed
+**This model has been trained for tool-use using the dr-agent-lib framework**.
+As such, running it out of the box with HuggingFace or vLLM will not work!
+Instead, you can run it like so:
 
-## Training and evaluation data
+```python
+TODO code snippet showing how to run model
+```
 
-More information needed
+See the [dr-agent-lib github](TODO) for more details on installation and how to run our model.
+Or check out our [live demo](TODO)!
 
-## Training procedure
+# Evaluation Results
 
-### Training hyperparameters
+We provide evaluation instructions in the [dr-agent-lib github](TODO).
+
+
+| Benchmark | SQAv2 | HealthBench | ResearchQA | DeepResearch Bench | SimpleQA | 2Wiki | WebWalker | Average |
+|:----------|:------:|:----------:|:---------:|:-------------------:|:------:|:-------:|-------:|-------:|
+| [Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) (naive rag; starting point model) | 40.4 | 16.5 | 56.1 | 33.3 | 52.6 | 18.9 | 8.8 | 32.4 |
+| [Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) (our pipeline; starting point model) | 57.2 | 5.9 | 46.3 | 18.2 | 70.5 | 44.0 | 27.9 | 38.6 |
+| [DR-Tulu-SFT-8B](https://huggingface.co/rl-research/DR-Tulu-SFT-8B) (**this model**) | 72.3 | 38.1 | 68.5 | 39.0 | 75.5 | 66.5 | 31.9 | 56.0 |
+| [DR-Tulu-8B](https://huggingface.co/rl-research/DR-Tulu-8B) | **86.7** | **43.7** |  **71.1** | **41.8** | **80.1** | **68.0** | **39.1** | **61.5** |
+
+For more baselines, explanations of this table, and analysis of reesults, check out the [Dr Tulu paper](https://arxiv.org/abs/TODO)!
+
+# Intended uses & limitations
+
+This model is licensed under Apache 2.0. It is intended for research and educational use in accordance with [Ai2's Responsible Use Guidelines](https://allenai.org/responsible-use).
+
+## Training
 
 The following hyperparameters were used during training:
 - learning_rate: 4e-05
@@ -49,13 +70,21 @@ The following hyperparameters were used during training:
 - lr_scheduler_warmup_ratio: 0.1
 - num_epochs: 5
 
-### Training results
-
+For futher details, check out the [Dr Tulu paper](https://arxiv.org/abs/TODO).
 
+# Links
+- 📝 [DR Tulu Paper](https://arxiv.org/abs/TODO)
+- ⚙️ [DR Tulu demo](https://dr-tulu.github.io/)
+- 💻 [DR Tulu code](https://github.com/rlresearch/DR-Tulu)
+- 🤖 [DR Tulu collection](https://huggingface.co/collections/rl-research/dr-tulu)
 
-### Framework versions
 
-- Transformers 4.52.4
-- Pytorch 2.8.0+cu128
-- Datasets 3.6.0
-- Tokenizers 0.21.1
+# Citation
+```
+@article{drtulu,
+  title = {{DR Tulu:  Reinforcement Learning with Evolving Rubrics for Deep Research}},
+  author = {{Rulin Shao, Akari Asai, Shannon Shen, Hamish Ivison, Varsha Kishore, Jingming Zhuo, Xinran Zhao, Molly Park, David Sontag, Tyler Murray, Sam Finlayson, Sewon Min, Pradeep Dasigi, Luca Soldani, Faeze Brahman, Scott Yih, Sherry Tongshuang Wu, Luke Zettlemoyer, Yoon Kim, Hanna Hajishirzi, Pang Wei Koh}},
+  journal={arXiv preprint TODO}
+  year = {2025},
+}
+```