Yong Zheng-Xin (Yong)
@yongzx.bsky.social
π€ 66
π₯ 80
π 14
ml research at Brown University // collab at Meta AI and Cohere For AI π yongzx.github.io
reposted by
Yong Zheng-Xin (Yong)
Cohere Labs
7 months ago
Itβs been two years since cross-lingual jailbreaks were first discovered. How far has the multilingual LLM safety research field advanced? π€ π Our comprehensive survey reveals that there is still a long way to go.
1
5
4
reposted by
Yong Zheng-Xin (Yong)
Julia Kreutzer
7 months ago
π¨LLM safety research needs to be at least as multilingual as our models. What's the current stage and how to progress from here? This work led by
@yongzx.bsky.social
has answers! π
add a skeleton here at some point
0
4
2
reposted by
Yong Zheng-Xin (Yong)
8 months ago
Can English-finetuned LLMs reason in other languages? Short Answer: Yes, thanks to βquote-and-thinkβ + test-time scaling. You can even force them to reason in a target language! But: π Low-resource langs & non-STEM topics still tough. New paper:
arxiv.org/abs/2505.05408
add a skeleton here at some point
1
2
1
reposted by
Yong Zheng-Xin (Yong)
Julia Kreutzer
8 months ago
Multilingual π€reasoning π€ test-time scaling π₯π₯π₯ New preprint!
@yongzx.bsky.social
has all the details π
add a skeleton here at some point
0
5
1
π£ New paper! We observe that reasoning language models finetuned only on English data are capable of zero-shot cross-lingual reasoning through a "quote-and-think" pattern. However, this does not mean they reason the same way across all languages or in new domains. [1/N]
8 months ago
2
11
5
you reached the end!!
feeds!
log in