nvidia/HelpSteer3
HelpSteer3 HelpSteer3 is an open-source dataset (CC-BY-4.0) that supports aligning models to become more helpful in responding to user prompts. HelpSteer3-Preference can be used to train Llama 3.3 Nemotron Super 49B v1 (for Generative RMs) and Llama 3.3 70B Instruct Models (for Bradley-Terry RMs) to produce Reward Models that score as high as 85.5% on RM-Bench and 78.6% on JudgeBench, which substantially surpass existing Reward Models on these benchmarks. HelpSteer3-Feedback and… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/HelpSteer3.
| Author / organization | nvidia |
|---|---|
| License (as declared) | cc-by-4.0 |
| Downloads | 14.3K (as reported by Hugging Face) |
| Likes | 120 |
| Created | 13 March 2025 |
| Last updated at source | 16 November 2025 |
Popularity, last 3 days
14,312 now (+3,380 since 2026-10-15). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- mradermacher/Q3-30B-A3B-Designant-i1-GGUFModel · used to train
- mradermacher/Q3-8B-Kintsugi-GGUFModel · used to train
- nvidia/Llama-3.3-Nemotron-70B-Reward-MultilingualModel · used to train
- opsd-genrm/SDRM-Qwen3-30B-A3BModel · used to train
- mradermacher/Q3-Nilcall-32B-v0.1-GGUFModel · used to train
- opsd-genrm/DR-GRPO-Qwen3-4BModel · used to train
- bartowski/allura-org_Q3-30B-A3B-Designant-GGUFModel · used to train
- mradermacher/Luminium-Gixel-Cube-v1-GGUFModel · used to train
- mradermacher/Llama-Poro-2-70B-Instruct-GGUFModel · used to train
- mradermacher/Llama-3.3-Nemotron-70B-Select-i1-GGUFModel · used to train
- bartowski/nvidia_Llama-3_3-Nemotron-Super-49B-GenRM-Multilingual-GGUFModel · used to train
- mradermacher/Llama-3.1-EstLLM-8B-Instruct-0825-GGUFModel · used to train
- opsd-genrm/DR-GRPO-Qwen3-30B-A3BModel · used to train
- mradermacher/Aitana-2B-S-Instruct-Aligned-GGUFModel · used to train
- mradermacher/Q3-Nilcall-8B-v0.3-GGUFModel · used to train
- tartuNLP/Llama-3.1-EstLLM-8B-Instruct-1125Model · used to train
- nvidia/Llama-3.3-Nemotron-70B-FeedbackModel · used to train
- jaredpalmer/kev-27bModel · used to train
- mradermacher/Role-mo-V2-32B-i1-GGUFModel · used to train
- mradermacher/MS3.2-Nilcall-24B-v0.1-i1-GGUFModel · used to train
- nvidia/Llama-3.3-Nemotron-70B-Reward-PrincipleModel · used to train
- nvidia/Llama-3_3-Nemotron-Super-49B-GenRM-MultilingualModel · used to train
- nvidia/Qwen-2.5-Nemotron-32B-RewardModel · used to train
- birgermoell/oellm-9b-threeway-dpo-step9000Model · used to train
- opsd-genrm/SDRM-Qwen3-4B-mask70Model · used to train
- mradermacher/Q3-Nilcall-8B-v0.2-GGUFModel · used to train
- mradermacher/Qwen3-Nemotron-14B-BRRM-GGUFModel · used to train
- mradermacher/Role-mo-V2-7B-GGUFModel · used to train
- mradermacher/Llama-3_3-Nemotron-Super-49B-GenRM-GGUFModel · used to train
- mradermacher/Qwen3-Nemotron-32B-GenRM-Principle-GGUFModel · used to train
- mradermacher/Luminium-Gixel-Cube-v1-i1-GGUFModel · used to train
- nvidia/Llama-3.3-Nemotron-70B-RewardModel · used to train
- mradermacher/Aitana-2B-SI-Instruct-Aligned-GGUFModel · used to train
- opsd-genrm/SDRM-Qwen3-30B-A3B-mask70Model · used to train
- mradermacher/Q3-Nilcall-32B-v0.3-i1-GGUFModel · used to train
- allura-org/Q3-30B-A3B-DesignantModel · used to train
- opsd-genrm/SDRM-Qwen3-4BModel · used to train
- nvidia/Qwen3-Nemotron-8B-BRRMModel · used to train
- mradermacher/Q3-Nilcall-8B-v0.3-i1-GGUFModel · used to train
- nvidia/Llama-3.3-Nemotron-70B-EditModel · used to train
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.