juanako-7b-UNA (Uniform Neural Alignment)

This model is a fine-tuned version of fblgit/juanako-7b-UNA-v2-phase-1 on the HuggingFaceH4/ultrafeedback_binarized dataset. It outperforms in many aspects most of the current Mistral based models and is the latest and most powerful juanako version as of now.

Scores

The official HuggingFace results can be found here

It scores: 59.91 according HuggingFace LLM Leaderboard. It scores: 65.1 with big-refactor branch of lm-eval-harness

Author Xavier M. @fblgit

Model description

juanako uses UNA, Uniform Neural Alignment. A training technique that ease alignment between transformer layers yet to be published.

Prompts

The following prompts showed positive results, it may depend the task and needs further experimentation but this should work for starters:

<|im_start|>system
- You are a helpful assistant chatbot trained by MosaicML.
- You answer questions.
- You are excited to be able to help the user, but will refuse to do anything that could be considered harmful to the user.
- You are more than just an information source, you are also able to write poetry, short stories, and make jokes.<|im_end|>
<|im_start|>user
Explain QKV<|im_end|>
<|im_start|>assistant

### Assistant: I am StableVicuna, a large language model created by CarperAI. I am here to chat!

### Human: Explain QKV
### Assistant:

[Round <|round|>]
问：Explain QKV
答：

[Round <|round|>]
Question：Explain QKV
Answer：

Question：Explain QKV
Answer：

Evaluations (lm-eval big-refactor branch)

TruthfulQA 0-Shot

| Tasks |Version|Filter|Metric|Value | |Stderr|
|--------------|-------|------|------|-----:|---|-----:|
|truthfulqa_mc2|Yaml |none |acc |0.6549|± |0.0153|

ARC 25-Shot

| Tasks |Version|Filter| Metric |Value | |Stderr|
|-------------|-------|------|--------|-----:|---|-----:|
|arc_challenge|Yaml |none |acc |0.6476|± |0.0140|
| | |none |acc_norm|0.6809|± |0.0136|

HellaSwag 10-Shot

| Tasks |Version|Filter| Metric |Value | |Stderr|
|---------|-------|------|--------|-----:|---|-----:|
|hellaswag|Yaml |none |acc |0.6703|± |0.0047|
| | |none |acc_norm|0.8520|± |0.0035|

GSM8k 5-Shot

|Tasks|Version| Filter | Metric |Value | |Stderr|
|-----|-------|----------|-----------|-----:|---|-----:|
|gsm8k|Yaml |get-answer|exact_match|0.4898|± |0.0138|

GPT Evaluations 0-Shot

| Tasks |Version|Filter| Metric |Value | |Stderr|
|--------------|-------|------|----------|-----:|---|-----:|
|boolq |Yaml |none |acc |0.8703|± |0.0059|
|lambada_openai|Yaml |none |perplexity|3.2598|± |0.0705|
| | |none |acc |0.7336|± |0.0062|
|piqa |Yaml |none |acc |0.8254|± |0.0089|
| | |none |acc_norm |0.8292|± |0.0088|
|sciq |Yaml |none |acc |0.9580|± |0.0063|
| | |none |acc_norm |0.9130|± |0.0089|

MathQA 0-Shot

|Tasks |Version|Filter| Metric |Value | |Stderr|
|------|-------|------|--------|-----:|---|-----:|
|mathqa|Yaml |none |acc |0.3752|± |0.0089|
| | |none |acc_norm|0.3772|± |0.0089|

PiQa 1-Shot

|Tasks|Version|Filter| Metric |Value | |Stderr|
|-----|-------|------|--------|-----:|---|-----:|
|piqa |Yaml |none |acc |0.8308|± |0.0087|
| | |none |acc_norm|0.8357|± |0.0086|

Winogrande 5-Shot

| Tasks |Version|Filter|Metric|Value| |Stderr|
|----------|-------|------|------|----:|---|-----:|
|winogrande|Yaml |none |acc |0.768|± |0.0119|

PubMedQA 0-Shot

| Tasks |Version|Filter|Metric|Value| |Stderr|
|--------|-------|------|------|----:|---|-----:|
|pubmedqa|Yaml |none |acc | 0.76|± |0.0191|

RACE 1-Shot

|Tasks|Version|Filter|Metric|Value | |Stderr|
|-----|-------|------|------|-----:|---|-----:|
|race |Yaml |none |acc |0.5282|± |0.0154|

MMLU 5-Shot (8-Bit)

| Groups |Version|Filter|Metric|Value | |Stderr|
|------------------|-------|------|------|-----:|---|-----:|
|mmlu |N/A |none |acc |0.6137|± |0.1243|
| - humanities |N/A |none |acc |0.5671|± |0.1101|
| - other |N/A |none |acc |0.6859|± |0.1164|
| - social_sciences|N/A |none |acc |0.7195|± |0.0713|
| - stem |N/A |none |acc |0.5087|± |0.1297|

DROP 3-Shot (8-Bit) (Instruct-Eval)

{'score': 0.49801113762927607}
{'drop': 49.8}
drop: 49.8

CRASS 0-Shot (Instruct-Eval)

{'score': 0.8357664233576643}
{'crass': 83.58}
crass: 83.58

Training Details

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 14
• gradient_accumulation_steps: 16
• total_train_batch_size: 224
• total_eval_batch_size: 14
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.01
• num_epochs: 1

Training results

Framework versions

• Transformers 4.35.0-UNA
• Pytorch 2.1.0
• Datasets 2.14.6
• Tokenizers 0.14.1

Citations

If you find juanako useful please:

@misc{juanako7buna,
 title={Juanako: Uniform Neural Alignment}, 
 author={Xavier Murias},
 year={2023},
 publisher = {HuggingFace},
 journal = {HuggingFace repository},
 howpublished = {\url{https://huggingface.co/fblgit/juanako-7b-UNA}},
}

Thanks to all the brilliant humans behind the creation of AI, here some of the ones that we find relevant to our research. If you feel a citation is missing, please contact.

@misc{lin2021truthfulqa,
 title={TruthfulQA: Measuring How Models Mimic Human Falsehoods},
 author={Stephanie Lin and Jacob Hilton and Owain Evans},
 year={2021},
 eprint={2109.07958},
 archivePrefix={arXiv},
 primaryClass={cs.CL}
}
@misc{tunstall2023zephyr,
 title={Zephyr: Direct Distillation of LM Alignment}, 
 author={Lewis Tunstall and Edward Beeching and Nathan Lambert and Nazneen Rajani and Kashif Rasul and Younes Belkada and Shengyi Huang and Leandro von Werra and Clémentine Fourrier and Nathan Habib and Nathan Sarrazin and Omar Sanseviero and Alexander M. Rush and Thomas Wolf},
 year={2023},
 eprint={2310.16944},
 archivePrefix={arXiv},
 primaryClass={cs.LG}
}
@inproceedings{Bisk2020,
 author = {Yonatan Bisk and Rowan Zellers and
 Ronan Le Bras and Jianfeng Gao
 and Yejin Choi},
 title = {PIQA: Reasoning about Physical Commonsense in
 Natural Language},
 booktitle = {Thirty-Fourth AAAI Conference on
 Artificial Intelligence},
 year = {2020},
}
@software{eval-harness,
 author = {Gao, Leo and
 Tow, Jonathan and
 Biderman, Stella and
 Black, Sid and
 DiPofi, Anthony and
 Foster, Charles and
 Golding, Laurence and
 Hsu, Jeffrey and
 McDonell, Kyle and
 Muennighoff, Niklas and
 Phang, Jason and
 Reynolds, Laria and
 Tang, Eric and
 Thite, Anish and
 Wang, Ben and
 Wang, Kevin and
 Zou, Andy},
 title = {A framework for few-shot language model evaluation},
 month = sep,
 year = 2021,
 publisher = {Zenodo},
 version = {v0.0.1},
 doi = {10.5281/zenodo.5371628},
 url = {https://doi.org/10.5281/zenodo.5371628}
}
@misc{rafailov2023direct,
 title={Direct Preference Optimization: Your Language Model is Secretly a Reward Model}, 
 author={Rafael Rafailov and Archit Sharma and Eric Mitchell and Stefano Ermon and Christopher D. Manning and Chelsea Finn},
 year={2023},
 eprint={2305.18290},
 archivePrefix={arXiv},
}

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

These are the Juanako 7B Trained with SFT & DDP & UNA • 8 items • Updated May 16 • 4