Update dataset card and add metadata for ViroBench

README.md

@@ -1,18 +1,81 @@
 ---
 license: cc-by-4.0
+task_categories:
+- text-classification
+- text-generation
+size_categories:
+- 10K<n<100K
 tags:
 - biology
-size_categories:
-- 100K<n<1M
+- genomics
+- virology
 ---
 
-# ViroBench Dataset (58,314 viral sequences)
+# ViroBench: Viral Understanding & Safety Benchmark for Nucleotide Foundation Models
 
 ViroBench is a curated viral genomics dataset and benchmark suite designed to **stress-test nucleotide foundation models (NFMs)** under realistic distribution shifts and to support **responsible viral sequence modeling**.
 
-- **Sequences:** 58,314 viral nucleotide sequences  
-- **Tasks:** 18 total (**12 classification + 6 generation**)  
-- **Project repo:** https://github.com/QIANJINYDX/ViroBench
+- **Paper:** [ViroBench: Benchmarking Nucleotide Foundation Models on Viral Genomics Tasks](https://huggingface.co/papers/2605.25388)
+- **Repository:** [https://github.com/QIANJINYDX/ViroBench](https://github.com/QIANJINYDX/ViroBench)
+- **Project Page / Leaderboard:** [https://sii-agi4s.github.io/ViroBench/](https://sii-agi4s.github.io/ViroBench/)
+
+## Dataset Summary
+
+ViroBench is a dual-axis diagnostic benchmark for viral sequence modeling, evaluating both:
+- **Understanding Axis**: Taxonomy classification & host prediction (12 classification tasks).
+- **Biosecurity Axis**: Genome modeling & CDS completion (6 generation tasks).
+
+The full dataset contains **58,314** viral samples curated from NCBI (GenBank + RefSeq), spanning DNA, RNA, and unknown nucleic-acid types. 
+
+### ViroBench-CLS-Lite
+A curated lightweight subset of **8,000** sequences designed for rapid prototyping, hyperparameter screening, and efficient evaluation of models on taxonomy and host classification tasks.
+
+## Benchmark Tasks
+
+### 1. Understanding Axis (Classification)
+- **Taxonomy Classification**: Prediction at Kingdom, Phylum, Class, Order, and Family levels.
+- **Host Prediction**: Predicting the virus host across different nucleic-acid types.
+- **Split Strategies**: Uses **Genus-disjoint** splits (to reduce phylogenetic leakage) and **Temporal** splits (to capture distribution drift).
+
+### 2. Biosecurity Axis (Generation)
+- **Genome Modeling**: Evaluation using bits-per-base (BPB) across length-bucketed tasks (short, medium, long).
+- **CDS Completion**: Evaluating coding sequence (CDS) continuation under functional constraints such as coding validity and frame consistency.
+
+## Sample Usage
+
+To run the benchmark using the official scripts, follow these examples from the repository:
+
+### Classification Benchmark
+```bash
+# Example: genus-disjoint taxonomy (all NA types)
+python script/run_all.py \
+  --model_name evo2_1b_base \
+  --dataset_name ALL-taxon-genus
+```
+
+### Generation Diagnostics (Genome Modeling)
+```bash
+python script/run_all_ppl.py \
+  --model_name evo2_1b_base \
+  --dataset_name genome-long
+```
+
+### CDS Completion
+```bash
+python script/run_all_gen.py \
+  --model_name evo2_1b_base \
+  --dataset_name cds-medium
+```
+
+## Citation
+
+If you find this benchmark useful, please cite the following:
 
-# ViroBench-CLS-Lite Dataset (8,000 viral sequences)
-ViroBench-CLS-Lite is a curated lightweight viral genomics subset designed to support rapid prototyping, hyperparameter screening, preliminary model comparison, and efficient evaluation of nucleotide foundation models on taxonomy and host classification tasks under realistic temporal and taxonomic distribution shifts.
+```bibtex
+@misc{virobench2025,
+      title={ViroBench: Benchmarking Nucleotide Foundation Models on Viral Genomics Tasks}, 
+      author={Qianjin Y. et al.},
+      year={2025},
+      url={https://huggingface.co/papers/2605.25388}
+}
+```