Llama 3.1 Nemotron Instruct 70B
NVIDIALlamaOpen WeightLlama 3.1 Community License
描述
A large language model customized by NVIDIA to improve the helpfulness of LLM generated responses. It is a fine-tuned version of Llama 3.1 70B Instruct. The model was trained using RLHF (REINFORCE) with HelpSteer2-Preference prompts.
发布日期
2024-10-15
参数规模
70.0B
上下文长度
131K
支持模态
text
能力雷达图
29
general
14
coding
27
reasoning
30
science估算
0
agents
0
multimodal
Science 在缺少专门科学评测时使用推理能力代理估算。
排行榜排名
基准测试分数 (LLM Stats)
Communication
MT-Bench
0.09 / 100自报
Finance
MMLU Chat
80.6%自报
MMLU
80.2%自报
TruthfulQA
58.6%自报
General
Instruct HumanEval
73.8%自报
ARC-C
69.2%自报
Language
Winogrande
84.5%自报
XLSum English
31.6%自报
Math
GSM8k
91.4%自报
GSM8K Chat
81.9%自报
Reasoning
HellaSwag
85.6%自报
AA 评测指数
Intelligence Index13.4
Math Index11.0
Coding Index10.8
Math 5000.7
Mmlu Pro0.7
Gpqa0.5
Ifbench0.3
Aime0.2
Scicode0.2
Tau20.2
Livecodebench0.2
Aime 250.1
Lcr0.1
Hle0.0
Terminalbench Hard0.0
LLM Stats 分类评分
Math90
Language80
Finance70
Healthcare70
Legal70
Reasoning70
General50
Communication10
Creativity10
Roleplay10
定价
输入价格$1.2 / 1M tokens
输出价格$1.2 / 1M tokens
混合价格(3:1)$1.2 / 1M tokens
速度
Tokens/秒38.1 tokens/s
首Token延迟0.34s
首回答延迟0.34s
可用提供商
(LS 内部计价单位)暂无提供商数据