Add library_name, arXiv metadata and project page link

Hi! I'm Niels, part of the community team at Hugging Face.

This PR improves the model card by:
- Adding `library_name: transformers` to the metadata, which enables the "Use in Transformers" button and automated code snippets.
- Adding the `arxiv` ID to the metadata to link the model with the paper [Surprisal-Guided Selection: Compute-Optimal Test-Time Strategies for Execution-Grounded Code Generation](https://huggingface.co/papers/2602.07670).
- Adding a link to the research project page in the header.

Everything else looks great!

README.md

@@ -1,44 +1,46 @@
 ---
-license: apache-2.0
 base_model: openai/gpt-oss-120b
-tags:
-  - gpu-kernel
-  - cuda
-  - code-generation
-  - reinforcement-learning
-  - grpo
-  - kernelbench
 datasets:
-  - ScalingIntelligence/KernelBench
+- ScalingIntelligence/KernelBench
 language:
-  - en
+- en
+license: apache-2.0
 pipeline_tag: text-generation
+library_name: transformers
+arxiv: 2602.07670
+tags:
+- gpu-kernel
+- cuda
+- code-generation
+- reinforcement-learning
+- grpo
+- kernelbench
 model-index:
-  - name: KernelBench-RLVR-120b
-    results:
-      - task:
-          type: text-generation
-          name: GPU Kernel Generation
-        dataset:
-          name: KernelBench L1
-          type: ScalingIntelligence/KernelBench
-        metrics:
-          - name: task_success_rate (K=64, 20 tasks)
-            type: custom
-            value: 90.0
-          - name: fast_1 (K=1, per-sample)
-            type: custom
-            value: 53.3
-          - name: correctness (training dist.)
-            type: accuracy
-            value: 98.4
+- name: KernelBench-RLVR-120b
+  results:
+  - task:
+      type: text-generation
+      name: GPU Kernel Generation
+    dataset:
+      name: KernelBench L1
+      type: ScalingIntelligence/KernelBench
+    metrics:
+    - type: custom
+      value: 90.0
+      name: task_success_rate (K=64, 20 tasks)
+    - type: custom
+      value: 53.3
+      name: fast_1 (K=1, per-sample)
+    - type: accuracy
+      value: 98.4
+      name: correctness (training dist.)
 ---
 
 # KernelBench-RLVR-120b
 
 A 120B-parameter model fine-tuned with GRPO (Group Relative Policy Optimization) for GPU kernel generation. This model was used to study compute-optimal test-time strategies in [Surprisal-Guided Selection](http://arxiv.org/abs/2602.07670), where we find that Best-of-N search with surprisal-guided selection recovers oracle performance at zero additional cost.
 
-**Paper**: [arXiv:2602.07670](http://arxiv.org/abs/2602.07670) | **Code**: [GitHub](https://github.com/jbarnes850/test-time-training)
+**Paper**: [arXiv:2602.07670](http://arxiv.org/abs/2602.07670) | **Project Page**: [Blog](https://jbarnes850.github.io/2026/02/02/surprisal-guided-selection/) | **Code**: [GitHub](https://github.com/jbarnes850/test-time-training)
 
 ## Quick Start
 
@@ -175,4 +177,4 @@ If you use this model, please cite [our paper](http://arxiv.org/abs/2602.07670):
 - [KernelBench](https://github.com/ScalingIntelligence/KernelBench) - Ouyang et al., 2025
 - [TTT-Discover](https://arxiv.org/abs/2601.16175) - Yuksekgonul et al., 2026
 - [SDPO](https://arxiv.org/abs/2601.20802) - Zeng et al., 2026
-- [Scalable Power Sampling](https://arxiv.org/abs/2601.21590) - Ji et al., 2026
+- [Scalable Power Sampling](https://arxiv.org/abs/2601.21590) - Ji et al., 2026