---
license: mit
base_model: ZhangShenao/SELM-Llama-3-8B-Instruct-iter-2
tags:
- alignment-handbook
- dpo
- trl
- selm
datasets:
- HuggingFaceH4/ultrafeedback_binarized
model-index:
- name: SELM-Llama-3-8B-Instruct-iter-3
  results: []
---



<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->



Self-Exploring Language Models: Active Preference Elicitation for Online Alignment.



# SELM-Llama-3-8B-Instruct-iter-3



This model is a fine-tuned version of [ZhangShenao/SELM-Llama-3-8B-Instruct-iter-2](https://huggingface.co/ZhangShenao/SELM-Llama-3-8B-Instruct-iter-2) using synthetic data based on on the HuggingFaceH4/ultrafeedback_binarized dataset.



## Model description



- Model type: A 8B parameter Llama3-based Self-Exploring Language Models (SELM).
- License: MIT



## Results



|                                        | AlpacaEval 2.0 (LC WR) | MT-Bench (Average) |
|----------------------------------------|------------------------|--------------------|
| [Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct)  |    &emsp; &emsp; &emsp;&emsp;         24.31         |  &emsp; &emsp; &emsp;         7.93       |
| [SELM-Llama-3-8B-Instruct-iter-1](https://huggingface.co/ZhangShenao/SELM-Llama-3-8B-Instruct-iter-1) |    &emsp; &emsp; &emsp;&emsp;         32.02         |  &emsp; &emsp; &emsp;         7.92       |
| [SELM-Llama-3-8B-Instruct-iter-2](https://huggingface.co/ZhangShenao/SELM-Llama-3-8B-Instruct-iter-2) |    &emsp; &emsp; &emsp;&emsp;         35.65         |  &emsp; &emsp; &emsp;         8.09       |
| [SELM-Llama-3-8B-Instruct-iter-3](https://huggingface.co/ZhangShenao/SELM-Llama-3-8B-Instruct-iter-3)  |    &emsp; &emsp; &emsp;&emsp;           33.47         |  &emsp; &emsp; &emsp;        8.29       |


### Training hyperparameters

The following hyperparameters were used during training:
- alpha: 0.0001
- beta: 0.01
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1

### Training results



### Framework versions

- Transformers 4.40.2
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.19.1