Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
configs:
|
3 |
+
- config_name: law_knowledge_prob
|
4 |
+
data_files:
|
5 |
+
- split: test
|
6 |
+
path: test.jsonl
|
7 |
+
task_categories:
|
8 |
+
- text-classification
|
9 |
+
- question-answering
|
10 |
+
- zero-shot-classification
|
11 |
+
language:
|
12 |
+
- en
|
13 |
+
tags:
|
14 |
+
- legal
|
15 |
+
---
|
16 |
+
|
17 |
+
# Domain Adaptation of Large Language Models
|
18 |
+
This repo contains the **Law Knowledge Probing dataset** used in our **ICLR 2024** paper [Adapting Large Language Models via Reading Comprehension](https://huggingface.co/papers/2309.09530).
|
19 |
+
|
20 |
+
We explore **continued pre-training on domain-specific corpora** for large language models. While this approach enriches LLMs with domain knowledge, it significantly hurts their prompting ability for question answering. Inspired by human learning via reading comprehension, we propose a simple method to **transform large-scale pre-training corpora into reading comprehension texts**, consistently improving prompting performance across tasks in biomedicine, finance, and law domains. **Our 7B model competes with much larger domain-specific models like BloombergGPT-50B**.
|
21 |
+
|
22 |
+
### 🤗 We are currently working hard on developing models across different domains, scales and architectures! Please stay tuned! 🤗
|
23 |
+
|
24 |
+
**************************** **Updates** ****************************
|
25 |
+
* 2024/4/14: Released the knowledge probing datasets at [med_knowledge_prob](https://huggingface.co/datasets/AdaptLLM/med_knowledge_prob) and [law_knowledge_prob](https://huggingface.co/datasets/AdaptLLM/law_knowledge_prob)
|
26 |
+
* 2024/4/2: Released the raw data splits (train and test) of all the evaluation datasets
|
27 |
+
* 2024/1/16: 🎉 Our [research paper](https://huggingface.co/papers/2309.09530) has been accepted by ICLR 2024!!!🎉
|
28 |
+
* 2023/12/19: Released our [13B base models](https://huggingface.co/AdaptLLM/law-LLM-13B) developed from LLaMA-1-13B.
|
29 |
+
* 2023/12/8: Released our [chat models](https://huggingface.co/AdaptLLM/law-chat) developed from LLaMA-2-Chat-7B.
|
30 |
+
* 2023/9/18: Released our [paper](https://huggingface.co/papers/2309.09530), [code](https://github.com/microsoft/LMOps), [data](https://huggingface.co/datasets/AdaptLLM/law-tasks), and [base models](https://huggingface.co/AdaptLLM/law-LLM) developed from LLaMA-1-7B.
|
31 |
+
|
32 |
+
|
33 |
+
## Domain-Specific LLMs
|
34 |
+
### LLaMA-1-7B
|
35 |
+
In our paper, we develop three domain-specific models from LLaMA-1-7B, which are also available in Huggingface: [Biomedicine-LLM](https://huggingface.co/AdaptLLM/medicine-LLM), [Finance-LLM](https://huggingface.co/AdaptLLM/finance-LLM) and [Law-LLM](https://huggingface.co/AdaptLLM/law-LLM), the performances of our AdaptLLM compared to other domain-specific LLMs are:
|
36 |
+
|
37 |
+
<p align='center'>
|
38 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/650801ced5578ef7e20b33d4/6efPwitFgy-pLTzvccdcP.png" width="700">
|
39 |
+
</p>
|
40 |
+
|
41 |
+
### LLaMA-1-13B
|
42 |
+
Moreover, we scale up our base model to LLaMA-1-13B to see if **our method is similarly effective for larger-scale models**, and the results are consistently positive too: [Biomedicine-LLM-13B](https://huggingface.co/AdaptLLM/medicine-LLM-13B), [Finance-LLM-13B](https://huggingface.co/AdaptLLM/finance-LLM-13B) and [Law-LLM-13B](https://huggingface.co/AdaptLLM/law-LLM-13B).
|
43 |
+
|
44 |
+
## Domain-Specific LLaMA-2-Chat
|
45 |
+
Our method is also effective for aligned models! LLaMA-2-Chat requires a [specific data format](https://huggingface.co/blog/llama2#how-to-prompt-llama-2), and our **reading comprehension can perfectly fit the data format** by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: [Biomedicine-Chat](https://huggingface.co/AdaptLLM/medicine-chat), [Finance-Chat](https://huggingface.co/AdaptLLM/finance-chat) and [Law-Chat](https://huggingface.co/AdaptLLM/law-chat)
|
46 |
+
|
47 |
+
## Domain-Specific Tasks
|
48 |
+
|
49 |
+
### Pre-templatized/Formatted Testing Splits
|
50 |
+
To easily reproduce our prompting results, we have uploaded the filled-in zero/few-shot input instructions and output completions of the test each domain-specific task: [biomedicine-tasks](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), [finance-tasks](https://huggingface.co/datasets/AdaptLLM/finance-tasks), and [law-tasks](https://huggingface.co/datasets/AdaptLLM/law-tasks).
|
51 |
+
|
52 |
+
**Note:** those filled-in instructions are specifically tailored for models before alignment and do NOT fit for the specific data format required for chat models.
|
53 |
+
|
54 |
+
### Raw Datasets
|
55 |
+
We have also uploaded the raw training and testing splits, for facilitating fine-tuning or other usages: [ChemProt](https://huggingface.co/datasets/AdaptLLM/ChemProt), [RCT](https://huggingface.co/datasets/AdaptLLM/RCT), [ConvFinQA](https://huggingface.co/datasets/AdaptLLM/ConvFinQA), [FiQA_SA](https://huggingface.co/datasets/AdaptLLM/FiQA_SA), [Headline](https://huggingface.co/datasets/AdaptLLM/Headline), [NER](https://huggingface.co/datasets/AdaptLLM/NER), [FPB](https://huggingface.co/datasets/AdaptLLM/FPB)
|
56 |
+
|
57 |
+
The other datasets used in our paper have already been available in huggingface.
|
58 |
+
|
59 |
+
### Domain Knowledge Probing
|
60 |
+
Our pre-processed knowledge probing datasets are available at: [med_knowledge_prob](https://huggingface.co/datasets/AdaptLLM/med_knowledge_prob) and [law_knowledge_prob](https://huggingface.co/datasets/AdaptLLM/law_knowledge_prob)
|
61 |
+
|
62 |
+
## Citation
|
63 |
+
If you find our work helpful, please cite us:
|
64 |
+
```bibtex
|
65 |
+
@inproceedings{
|
66 |
+
cheng2024adapting,
|
67 |
+
title={Adapting Large Language Models via Reading Comprehension},
|
68 |
+
author={Daixuan Cheng and Shaohan Huang and Furu Wei},
|
69 |
+
booktitle={The Twelfth International Conference on Learning Representations},
|
70 |
+
year={2024},
|
71 |
+
url={https://openreview.net/forum?id=y886UXPEZ0}
|
72 |
+
}
|
73 |
+
```
|
74 |
+
|
75 |
+
and the original dataset:
|
76 |
+
```bibtex
|
77 |
+
@inproceedings{LEDGAR,
|
78 |
+
author = {Don Tuggener and
|
79 |
+
Pius von D{\"{a}}niken and
|
80 |
+
Thomas Peetz and
|
81 |
+
Mark Cieliebak},
|
82 |
+
title = {{LEDGAR:} {A} Large-Scale Multi-label Corpus for Text Classification
|
83 |
+
of Legal Provisions in Contracts},
|
84 |
+
booktitle = {{LREC}},
|
85 |
+
pages = {1235--1241},
|
86 |
+
publisher = {European Language Resources Association},
|
87 |
+
year = {2020}
|
88 |
+
}
|
89 |
+
```
|