Statements (43)
Predicate | Object |
---|---|
gptkbp:instanceOf |
large language model
|
gptkbp:architecture |
gptkb:transformation
|
gptkbp:availableOn |
gptkb:Hugging_Face
gptkb:Meta_AI_website |
gptkbp:benchmarkPerformance |
competitive with GPT-3.5 on some tasks
|
gptkbp:citation |
gptkb:Touvron_et_al.,_2023
|
gptkbp:developedBy |
gptkb:Meta_AI
|
gptkbp:documentation |
Llama 2 technical report
|
gptkbp:fineTunedWith |
gptkb:public_speaker
|
gptkbp:hasModel |
decoder-only transformer
|
https://www.w3.org/2000/01/rdf-schema#label |
Llama 2 13B Chat
|
gptkbp:input |
4096 tokens
|
gptkbp:intendedUse |
conversational AI
|
gptkbp:license |
custom commercial license
|
gptkbp:notableFeature |
instruction following
longer context window improved safety over LLaMA |
gptkbp:notableFor |
research
prototyping AI applications enterprise chatbots |
gptkbp:openSource |
partially
|
gptkbp:optimizedFor |
chat applications
|
gptkbp:parameter |
13B
|
gptkbp:partOf |
Llama 2 family
|
gptkbp:platform |
GPU with at least 24GB VRAM (for inference)
|
gptkbp:predecessor |
gptkb:LLaMA
|
gptkbp:releaseDate |
July 2023
|
gptkbp:size |
13 billion parameters
|
gptkbp:successor |
Llama 3 (for some use cases)
|
gptkbp:supports |
English
code generation question answering summarization text generation reasoning tasks |
gptkbp:trainer |
Meta-licensed data
data from partners supervised fine-tuning publicly available data sources reinforcement learning from human feedback (RLHF) |
gptkbp:weight |
yes (with restrictions)
|
gptkbp:bfsParent |
gptkb:LLaMA_model_family
|
gptkbp:bfsLayer |
8
|