Add GitHub link; paper and citation coming soon

README.md

@@ -16,8 +16,8 @@ per-dataset training.
 
 This variant was post-trained with expression-size penalty λ = 0.15. It trades a little fit for smaller expressions than the default RASR-94M-L.
 
-- **Paper:** *Improving Neural Symbolic Regression with a Diverse Prior and Post-training Alignment* (TODO: link)
-- **Code:** TODO: GitHub link
+- **Paper:** *Improving Neural Symbolic Regression with a Diverse Prior and Post-training Alignment* (coming soon)
+- **Code:** [github.com/automl/rasr](https://github.com/automl/rasr)
 
 ## Variants
 
@@ -30,7 +30,7 @@ This variant was post-trained with expression-size penalty λ = 0.15. It trades
 
 ## Usage
 
-Install the RASR code (TODO: GitHub link). The weights are downloaded automatically on first use:
+Install the RASR code from [github.com/automl/rasr](https://github.com/automl/rasr). The weights are downloaded automatically on first use:
 
 ```python
 from src.rasr.inference.rasr import RASR
@@ -66,6 +66,4 @@ See the paper for the full evaluation.
 
 ## Citation
 
-```bibtex
-TODO
-```
+Coming soon.