Jirui Qi
@Jirui_Qi
Followers
131
Following
140
Statuses
47
Ph.D Candidate @GroNLP, University of Groningen #NLProc
Groningen, The Netherlands
Joined June 2023
🚨 Our paper on **efficient prompt engineering** has been accepted by NAACL2025 Main Conference! Key Point: LLMs tend to generate better responses when the likelihood of the question segment is higher. I.e. p(question) ∝ Performance Paper available at:
[1/8] Seeking a faster approach to assess prompt quality in RAG QA? Our latest work may be a good fit for you. We find that prompt quality can be measured using question likelihoods, a computation that’s parallelizable on the input side of LLMs! 📄 #NLProc
0
5
27
RT @yuji_amanogawa: RAGの回答精度と正解の位置の関係はよく話題になりますが、「そもそも正解文を特定できれば苦労しないよ」という問題への解決策になりそうな興味深い論文ですね。…
0
21
0
@yuji_amanogawa Thank you for your interest and the sharing of our latest work! 論文に関心を持っていただき、ありがとうございます！
1
0
1
Memorable experience @emnlpmeeting in Miami. It is interesting to present our latest work "Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation" and exchange ideas with amazing people. Can't wait to see you in Suzhou next year! #NLProc #NLP
0
2
24
Many thanks to all collaborators for their contributions! @t_y_liu, Paul He, @AriannaBisazza, @mrinmayasachan, Ryan Cotterell.
0
0
4
RT @xchen2023: 🧠 Can LLMs actually reason? We have surprising observations on simple sequential reasoning tasks! Our paper in #EMNLP2024 Fi…
0
4
0