Skip to content

Official Repository for SafeDecoding: Defending against Jailbreak Attacks via Safety-Aware Decoding

active 2024-02-182026-02-10 (UTC)

Partial coverage20,502 / 26,258 hourly files (78%) · 2 absent upstream · 5,753 failed, retryable2023-08-152026-08-13 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
151
Pushes
19
Pull requests
0
Issues
11
Stars
86
Forks
14

Activity over time

Daily event counts in the loaded window

Line chart, 724 days from 2024-02-18 to 2026-02-10. Pushes: 19 total, peak 6 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 11 total, peak 1 in a day. Comments: 16 total, peak 2 in a day. Stars: 86 total, peak 3 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

ContributorContributionsPushesPRsComments
fly-dust281806
SCccc213002
terarachang3002
zhangchen-xu2001
Renpf20222001
shanpoyang6542002
Deep19941000
LetheSec1001
Django-Jiang1100
xunguangwang1001
ZHANGJINKUI1000
chenzongxiong1000

Recent activity

Latest issues, pull requests and releases

  • Issue#11ZHANGJINKUI2025-03-30 02:57
    lora_modules
  • Issue#10zhangchen-xu2025-02-08 08:34
    Question about the "ATGR"
  • Issue comment#10zhangchen-xu2025-01-11 22:38
    Question about the "ATGR"
  • Issue comment#9xunguangwang2025-01-11 10:03
    Dataset Format
  • Issue#10Deep19942024-12-26 12:47
    Question about the "ATGR"
  • Issue#7fly-dust2024-10-18 07:39
    code for AutoDAN, GCG, DeepInception and PAIR attacks
  • Issue comment#7fly-dust2024-09-15 00:41
    code for AutoDAN, GCG, DeepInception and PAIR attacks
  • Issue#7chenzongxiong2024-09-14 08:06
    code for AutoDAN, GCG, DeepInception and PAIR attacks
  • Issue#6fly-dust2024-07-19 06:06
    Doubts about the ft_datasets format
  • Issue comment#6fly-dust2024-07-14 18:51
    Doubts about the ft_datasets format
  • Issue comment#6Renpf20222024-07-14 11:51
    Doubts about the ft_datasets format
  • Issue#6Renpf20222024-07-14 09:48
    Doubts about the ft_datasets format
  • Issue comment#5fly-dust2024-07-09 06:11
    outputs of Expert models start with some special tokens while the ones of Base models do not
  • Issue comment#5terarachang2024-07-05 20:09
    outputs of Expert models start with some special tokens while the ones of Base models do not
  • Issue comment#5terarachang2024-07-05 19:57
    outputs of Expert models start with some special tokens while the ones of Base models do not
  • Issue comment#5fly-dust2024-07-04 20:18
    outputs of Expert models start with some special tokens while the ones of Base models do not
  • Issue#5terarachang2024-07-04 00:37
    outputs of Expert models start with some special tokens while the ones of Base models do not
  • Issue comment#4SCccc212024-05-22 03:19
    Dataset used for finetuning the expert model
  • Issue#4SCccc212024-05-22 03:19
    Dataset used for finetuning the expert model
  • Issue comment#4fly-dust2024-05-17 04:50
    Dataset used for finetuning the expert model
  • Issue comment#4SCccc212024-05-14 02:20
    Dataset used for finetuning the expert model
  • Issue comment#4fly-dust2024-05-12 06:02
    Dataset used for finetuning the expert model
  • Issue comment#3shanpoyang6542024-04-03 06:49
    why is `output_expert` the output of the expert model?
  • Issue comment#3shanpoyang6542024-04-03 06:46
    why is `output_expert` the output of the expert model?
  • Issue#2fly-dust2024-03-08 06:47
    About adaptive attack

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 86 stars here means stars gained during the window, not the repo's star count.