nvidia/Nemotron-ClimbMix
ClimbMix Dataset π Creating the highest-quality pre-training datasets for LLMs π π PAPER π€ CLIMBLAB π€ CLIMBMIX π HOMEPAGE Figure 1: Continuously training a 1B model yields a 2.0% improvement over Llama-3.2-1B, demonstrating a more efficient scaling trend compared to prior models. Figure 2: Pre-training a 1B model from scratch on ClimbMix shows better scaling effects than training on otherβ¦ See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Nemotron-ClimbMix.
| Author / organization | nvidia |
|---|---|
| License (as declared) | cc-by-nc-4.0 |
| Downloads | 7.5K (as reported by Hugging Face) |
| Likes | 132 |
| Created | 26 March 2025 |
| Last updated at source | 21 October 2025 |
Popularity, last 2 days
7,498 now (-8 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- qikp/kite-8.1-11m-baseModel Β· used to train
- MinimaLabs/min-spark-1.6Model Β· used to train
- AxiomicLabs/GPT-S3-3MModel Β· used to train
- AxiomicLabs/GPT-X3-150MModel Β· used to train
Recent changes
- 17 Oct 2026 Description: ClimbMix Dataset π Creating the highest-quality pre-training datasets for LLMs π π PAPER π€ CLIMBLAB π€ CLIMBMIX π HOMEPAGE Figure 1: Continuously training a 1B model yields a 2.0% improvement over Llama-3.2-1B, demonstrating a more eff β ClimbMix Dataset π Creating the highest-quality pre-training datasets for LLMs π π PAPER π€ CLIMBLAB π€ CLIMBMIX π HOMEPAGE Figure 1: Continuously training a 1B model yields a 2.0% improvement over Llama-3.2-1B, demonstrating a more eff
Changes detected by Silytics between refreshes of the source.
More from nvidia
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.