Update broken links

README.md

@@ -87,11 +87,11 @@ Every model NVIDIA ships rests on a data layer — and that data shapes how the
 | **FOUNDATION** |  |  |
 | Pre-train a base model | [**Nemotron Pre-Training Collection**](https://huggingface.co/collections/nvidia/nemotron-pre-training-datasets) | Nemotron-CC-v2.1, Nemotron-CC-Math-v1, Nemotron-CC-Code-v1, Nemotron-ClimbMix |
 | **BUILD A CAPABILITY** |  |  |
-| Math reasoning, proofs, and quantitative problem-solving | [**Nemotron Math & Reasoning Collection**](https://huggingface.co/collections/nvidia/nemotron-math-reasoning) | Nemotron-SFT-Math-v3, Nemotron-Math-v2, AceReason-Math, Nemotron-CC-Math-v1 |
-| Code generation, debugging, and SWE workflows | [**Nemotron Code & SWE Collection**](https://huggingface.co/collections/nvidia/nemotron-code-swe) | Nemotron-SFT-Competitive-Programming-v2, Nemotron-SFT-SWE-v2, Nemotron-CC-Code-v1 |
-| Helpful, multi-turn, instruction-following chat | [**Nemotron Chat & Instruction Collection**](https://huggingface.co/collections/nvidia/nemotron-chat-instruction-following) | Nemotron-SFT-Instruction-Following-Chat-v2, Nemotron-RL-instruction\_following |
-| Agentic and tool-use behavior | [**Nemotron Agentic Collection**](https://huggingface.co/collections/nvidia/nemotron-agentic-tool-use) | Nemotron-SFT-Agentic-v2, Nemotron-RL-Agentic-Function-Calling-Pivot-v1 |
-| Safety, refusals, and content moderation | [**Nemotron Safety Collection**](https://huggingface.co/collections/nvidia/nemotron-safety-content-moderation) | Aegis-AI-Content-Safety-Dataset-2.0, Nemotron-Safety-Guard-Dataset-v3, Nemotron-PII |
+| Math reasoning, proofs, and quantitative problem-solving | [**Nemotron Math & Reasoning Collection**](https://huggingface.co/collections/nvidia/nemotron-math-and-reasoning) | Nemotron-SFT-Math-v3, Nemotron-Math-v2, AceReason-Math, Nemotron-CC-Math-v1 |
+| Code generation, debugging, and SWE workflows | [**Nemotron Code & SWE Collection**](https://huggingface.co/collections/nvidia/nemotron-code-and-swe) | Nemotron-SFT-Competitive-Programming-v2, Nemotron-SFT-SWE-v2, Nemotron-CC-Code-v1 |
+| Helpful, multi-turn, instruction-following chat | [**Nemotron Chat & Instruction Collection**](https://huggingface.co/collections/nvidia/nemotron-chat-and-instruction-following) | Nemotron-SFT-Instruction-Following-Chat-v2, Nemotron-RL-instruction\_following |
+| Agentic and tool-use behavior | [**Nemotron Agentic Collection**](https://huggingface.co/collections/nvidia/nemotron-agentic-and-tool-use) | Nemotron-SFT-Agentic-v2, Nemotron-RL-Agentic-Function-Calling-Pivot-v1 |
+| Safety, refusals, and content moderation | [**Nemotron Safety Collection**](https://huggingface.co/collections/nvidia/nemotron-safety-and-content-moderation) | Aegis-AI-Content-Safety-Dataset-2.0, Nemotron-Safety-Guard-Dataset-v3, Nemotron-PII |
 | Image and document understanding | [**Nemotron Vision-Language Collection**](https://huggingface.co/collections/nvidia/nemotron-vision-language) | Nemotron-VLM-Dataset-v2, Llama-Nemotron-VLM-Dataset-v1 |
 | **TRAINING STAGES** |  |  |
 | RL data with verifiable rewards (math, code, agentic, instruction) | [**Nemotron Reinforcement Learning Collection**](https://huggingface.co/collections/nvidia/nemotron-reinforcement-learning) | Nemotron-RL-math-OpenMathReasoning, Nemotron-RL-coding-competitive\_coding, Nemotron-RL-Agentic-Function-Calling-Pivot-v1 |