allenhzy commited on
Commit
f6cc227
1 Parent(s): 8845aa6
Files changed (1) hide show
  1. index.html +48 -1
index.html CHANGED
@@ -26,6 +26,19 @@
26
  <script src="./static/js/bulma-slider.min.js"></script>
27
  <script src="./static/js/index.js"></script>
28
 
 
 
 
 
 
 
 
 
 
 
 
 
 
29
  <style type="text/css">
30
  .tg {border-collapse:collapse;border-spacing:0;}
31
  .tg td{border-color:black;border-style:solid;border-width:1px;font-family:Arial, sans-serif;font-size:14px;
@@ -397,7 +410,41 @@
397
  <h2 class="title is-3">Adaptive Attack</h2>
398
  <div class="columns is-centered">
399
  <div class="column container-centered">
400
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
401
  </div>
402
  </div>
403
 
 
26
  <script src="./static/js/bulma-slider.min.js"></script>
27
  <script src="./static/js/index.js"></script>
28
 
29
+ <script>
30
+ $('#adaptive-loss-formula-list').on('click', 'a', function(e) {
31
+ e.preventDefault();
32
+ if (!$(this).hasClass('selected')) {
33
+ $('.formula').hide(200);
34
+ $('.formula-list > a').removeClass('selected');
35
+ $(this).addClass('selected');
36
+ var target = $(this).attr('href');
37
+ $(target).show(200);
38
+ }
39
+ });
40
+ </script>
41
+
42
  <style type="text/css">
43
  .tg {border-collapse:collapse;border-spacing:0;}
44
  .tg td{border-color:black;border-style:solid;border-width:1px;font-family:Arial, sans-serif;font-size:14px;
 
410
  <h2 class="title is-3">Adaptive Attack</h2>
411
  <div class="columns is-centered">
412
  <div class="column container-centered">
413
+ <div id="refusal-loss-formula" class="container">
414
+ <div id="adaptive-loss-formula-list" class="row align-items-center formula-list">
415
+ <a href="#label-loss" class="selected">Label Loss</a>
416
+ <a href="#representation-loss">Representation Loss</a>
417
+ <a href="#total-loss">Total Loss</a>
418
+ <div style="clear: both"></div>
419
+ </div>
420
+ <div id="adaptive" class="row align-items-center">
421
+ <span id="label-loss" class="formula" style="">
422
+ $$
423
+ \displaystyle
424
+ \begin{aligned}
425
+ \phi_\theta(x)&=1-\mathbb{E}_{y \sim T_\theta(x)} JB(y)\\
426
+ JB (y) &= \begin{cases}
427
+ 1 \text{, if $y$ contains any jailbreak keyword;} \\
428
+ 0 \text{, otherwise.}
429
+ \end{cases}
430
+ \end{aligned}
431
+ $$
432
+ </span>
433
+ <span id="representation-loss" class="formula" style="display: none;">
434
+ $$
435
+ \displaystyle
436
+ \begin{aligned}
437
+ f_\theta(x) &=1-\frac{1}{N}\sum_{i=1}^N JB(y_i)\\
438
+ JB (y_i) &= \begin{cases}
439
+ 1 \text{, if $y_i$ contains any jailbreak keyword;} \\
440
+ 0 \text{, otherwise.}
441
+ \end{cases}
442
+ \end{aligned}
443
+ $$
444
+ </span>
445
+ <span id="total-loss" class="formula" style="display: none;">$$\displaystyle g_\theta(x)=\sum_{i=1}^P \frac{f_\theta(x\oplus \mu u_i)-f_\theta(x)}{\mu} u_i $$</span>
446
+ </div>
447
+ </div>
448
  </div>
449
  </div>
450