-
Notifications
You must be signed in to change notification settings - Fork 7
/
Copy pathindex.html
354 lines (313 loc) · 15.3 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="Reformatted Alignment">
<meta name="keywords" content="Alignment, Language Models, NLP">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Reformatted Alignment</title>
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-PYVRSFMDRL"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() {
dataLayer.push(arguments);
}
gtag('js', new Date());
gtag('config', 'G-PYVRSFMDRL');
</script>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="./static/css/bulma.min.css">
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./static/css/bulma-slider.min.css">
<link rel="stylesheet" href="./static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="./static/css/index.css">
<link rel="icon" href="./static/images/mathpile-logo-1.png">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./static/js/fontawesome.all.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">Reformatted Alignment</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://scholar.google.com/citations?user=mhot7AUAAAAJ&hl=en" target="_blank">Run-Ze Fan</a><sup>1,6</sup>,</span>
<span class="author-block">
<a href="https://github.com/hongtangshui" target="_blank">Xuefeng Li</a><sup>1,6</sup>,</span>
<span class="author-block">
<a href="https://openreview.net/profile?id=~Haoyang_Zou1" target="_blank">Haoyang Zou</a><sup>3,6</sup>,</span>
<span class="author-block">
<a href="https://lockon-n.github.io/" target="_blank">Junlong Li</a><sup>1,6</sup>,</span>
<span class="author-block">
<a href="https://shwai-he.github.io/" target="_blank">Shwai He</a><sup>4</sup>,</span>
<span class="author-block">
<a href="https://ethanc111.github.io/" target="_blank">Ethan Chern</a><sup>1,6</sup>,</span>
<span class="author-block">
<a href="https://www.linkedin.com/in/jiewen-hu/" target="_blank">Jiewen Hu</a><sup>5,6</sup>,</span>
<span class="author-block">
<a href="http://pfliu.com/" target="_blank">Pengfei Liu</a><sup>1,2,6*</sup>
</span>
<!-- <span class="author-block">
<a href="http://sofienbouaziz.com">Sofien Bouaziz</a><sup>2</sup>,
</span>
<span class="author-block">
<a href="https://www.danbgoldman.com">Dan B Goldman</a><sup>2</sup>,
</span>
<span class="author-block">
<a href="https://homes.cs.washington.edu/~seitz/">Steven M. Seitz</a><sup>1,2</sup>,
</span>
<span class="author-block">
<a href="http://www.ricardomartinbrualla.com">Ricardo Martin-Brualla</a><sup>2</sup>
</span> -->
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>Shanghai Jiao Tong University,</span>
<span class="author-block"><sup>2</sup>Shanghai Artificial Intelligence Laboratory,</span>
<span class="author-block"><sup>3</sup>Fudan University,</span>
<span class="author-block"><sup>4</sup>University of Maryland, College Park,</span>
<span class="author-block"><sup>5</sup>CMU,</span>
<span class="author-block"><sup>6</sup>Generative AI Research Lab (GAIR)</span>
<span class="author-block"><sup>*</sup>Corresponding Author</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<!-- <span class="link-block">-->
<!-- <a href="https://huggingface.co/papers/2312.17120" target="_blank"-->
<!-- class="external-link button is-normal is-rounded is-dark">-->
<!-- <span class="icon">-->
<!-- <i class="fas fa-file-pdf"></i>-->
<!-- </span>-->
<!-- <span>Paper</span>-->
<!-- </a>-->
<!-- </span>-->
<span class="link-block">
<a href="https://arxiv.org/abs/2402.12219" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>Paper</span>
</a>
</span>
<!-- Video Link. -->
<!-- <span class="link-block">
<a href="https://www.youtube.com/watch?v=MrKrnHhk8IA"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-youtube"></i>
</span>
<span>Video</span>
</a>
</span> -->
<!-- Code Link. -->
<span class="link-block">
<a href="https://github.com/GAIR-NLP/ReAlign" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Code</span>
</a>
</span>
<!-- Dataset Link. -->
<!-- <span class="link-block">-->
<!-- <a href="https://huggingface.co/datasets/GAIR/MathPile" target="_blank"-->
<!-- class="external-link button is-normal is-rounded is-dark">-->
<!-- <span class="icon">-->
<!-- <i class="far fa-images"></i>-->
<!-- </span>-->
<!-- <span>Datasets</span>-->
<!-- </a>-->
<!-- </span>-->
<!-- Twitter Link. -->
<!-- <span class="link-block">-->
<!-- <a href="https://twitter.com/_akhaliq/status/1740571256234057798" target="_blank"-->
<!-- class="external-link button is-normal is-rounded is-dark">-->
<!-- <span class="icon">-->
<!-- <i class="fab fa-twitter"></i>-->
<!-- </span>-->
<!-- <span>Twitter by AK</span>-->
<!-- </a>-->
<!-- </span>-->
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Abstract. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Introduction</h2>
<div class="content has-text-justified">
<p>
We explores elevating the quality of existing instruction data to better align with human values, introducing a simple and effective approach named <b>ReAlign</b> (<b>Re</b>formatted <b>Align</b>ment), which <i>reformats</i> the responses of instruction data into a format that better aligns with pre-established criteria and the collated evidence.
This approach minimizes human annotation, hallucination, and the difficulty in scaling, remaining orthogonal to existing alignment techniques.
Experimentally, ReAlign significantly boosts the general alignment ability, math reasoning, factuality, and readability of the LLMs.
</p>
<p>
Encouragingly, <i>without</i> introducing any additional data or advanced training techniques, and merely by reformatting the response, LLaMA-2-13B's mathematical reasoning ability on GSM8K can be improved <b>from 46.77% to 56.63%</b> in accuracy.
Additionally, a mere 5% of ReAlign data yields a 67% boost in general alignment ability measured by the Alpaca dataset.
This work highlights the need for further research into the <i>science</i> and <i>interpretability</i> of LLMs.
</p>
<p>
The underlying <i>philosophy</i> of ReAlign is to re-coordinate the roles of humans and LLMs in the alignment process, leveraging their complementary strengths -- humans articulate their preferences, and LLMs, in turn, reconstruct instructions based on their generative power (e.g., instruction-following ability), without directly using distilled LLM knowledge.
Through this collaborative synergy, we expect the generated instruction data to be not only more contextually precise but also more closely aligned with human preferences.
</p>
<figure>
<img src="figs/Math_Results.jpg" width=100% alt="Description of first image"/>
<figcaption>
The accuracy of the GSM8K test set for LLaMA-2-13B and Mistral-7B models fine-tuned on the training set of GSM8K and MATH with and without <b>ReAlign</b>. (a): Training and testing on GSM8K. (b): Training on MATH and testing on GSM8K (Out-of-Distribution Setting).
</figcaption>
</figure>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Methodology</h2>
<div class="content has-text-justified">
<figure>
<img src="figs/overall_figs.jpg" width=100% alt="Description of second image"/>
<figcaption>
An overview of our <strong>ReAlign</strong> including three steps. KILT denotes Knowledge Intensive Language Tasks.
</figcaption>
</figure>
<p>The ReAlign process unfolds in three main steps. </p>
<p>The first step involves <b>criteria definition</b>, where humans define their preferences (e.g., the preferred format of responses) in various scenarios in the form of natural language.
In this paper, we meticulously define criteria for 46 distinct scenarios. </p>
<p>The second step, <b>retrieval augmentation</b>, broadens the knowledge base for knowledge-intensive tasks like open-domain QA and fact verification. This is achieved by incorporating additional information, thereby improving the factuality and informativeness of responses. </p>
<p>The final step, <b>reformatting</b>, aims to re-align the responses with the pre-established criteria and the collated evidence, guaranteeing outputs that are both structured and substantiated.</p>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Examples</h2>
<div class="content has-text-justified">
<figure>
<img src="figs/intro_graph.jpg" width=80% alt="Description of third image"/>
<figcaption>
<b>ReAlign</b> realigns the original response with the pre-defined criteria to be a better format.
</figcaption>
</figure>
<figure>
<img src="figs/model_example.jpg" width=100% alt="Description of fourth image"/>
<figcaption>
An example of the response from the original model and the response from the <b>ReAlign</b> Model.
</figcaption>
</figure>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Results</h2>
<div class="content has-text-justified">
<figure>
<img src="figs/alignment_results.jpg" width=100% alt="Description of fifth image"/>
<figcaption>
The results of the general alignment ability on the original datasets and the ReAlign datasets.
</figcaption>
</figure>
<figure>
<img src="figs/math_reasoning.jpg" width=60% alt="Description of sixth image"/>
<figcaption>
The results of math reasoning on GSM8K, MATH and them + ReAlign based on LLaMA-2-13B and Mistral-7B. We test models on both GSM8K and MATH test sets. We report the accuracy by exact matching.
</figcaption>
</figure>
<figure>
<img src="figs/factuality_results.jpg" width=60% alt="Description of seventh image"/>
<figcaption>
The results of the factuality score.
</figcaption>
</figure>
<figure>
<img src="figs/readability_results.jpg" width=100% alt="Description of eighth image"/>
<figcaption>
The readability win-rate of the original dataset + ReAlign against the original dataset based on LLaMA-2-13B, judged by GPT-4 and human.
</figcaption>
</figure>
<figure>
<img src="figs/scaling_law.jpg" width=80% alt="Description of ninth image"/>
<figcaption>
The scaling trends in ReAlign data percentage, including general alignment ability and knowledge ability. We conduct the experiment in the Alpaca dataset based on LLaMA-2-13B.
</figcaption>
</figure>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@article{fan2024reformatted,
title={Reformatted Alignment},
author={Fan, Run-Ze and Li, Xuefeng and Zou, Haoyang and Li, Junlong and He, Shwai and Chern, Ethan and Hu, Jiewen and Liu, Pengfei},
year={2024},
journal={arXiv preprint arXiv:2402.12219},
url={https://arxiv.org/abs/2402.12219}
}</code></pre>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link"
href="https://arxiv.org/abs/2402.12219">
<i class="fas fa-file-pdf"></i>
</a>
<a class="icon-link" href="https://github.com/GAIR-NLP/ReAlign" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div>
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This website is licensed under a <a rel="license"
href="http://creativecommons.org/licenses/by-sa/4.0/">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
<p>
This means you are free to borrow the <a
href="https://github.com/nerfies/nerfies.github.io">source code</a> of this website,
we just ask that you link back to this page in the footer.
Please remember to remove the analytics code included in the header of the website which
you do not want on your website.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>