OrpoLlama-3-8B

Home/Models/mlabonne/OrpoLlama-3-8B

mlabonne/OrpoLlama-3-8B

kind: original
quant: bf16 (16-bit)
family: llama
modality: text-generation
root model: meta-llama/Meta-Llama-3-8B
provider: mlabonne
license: other
language: en
context: —
library: transformers
vRAM (GB): —
size on disk: —
downloads: 45
likes: 53
created: —
last modified: 2024-06-15T14:01:22+00:00 (2024)
huggingface: https://huggingface.co/mlabonne/OrpoLlama-3-8B
tags: transformerssafetensorsllamatext-generationorpollama 3rlhfsftconversationalendataset:mlabonne/orpo-dpo-mix-40kbase_model:meta-llama/Meta-Llama-3-8Bbase_model:finetune:meta-llama/Meta-Llama-3-8Blicense:othertext-generation-inferenceendpoints_compatibleregion:us

benchmark	category	score	conf	evidence
TruthfulQA	safety	51.65	low	`\| [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) [📄](https://gist.github.com/mlabonne/8329284d86035e6019edb11eb0933628) \| 51.34 \|`
BigBench	reasoning	42.64	low	`\| [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) [📄](https://gist.github.com/mlabonne/8329284d86035e6019edb11eb0933628) \| 51.34 \|`
AGIEval	reasoning	41.22	low	`\| [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) [📄](https://gist.github.com/mlabonne/8329284d86035e6019edb11eb0933628) \| 51.34 \|`

none scraped.