Established 1993, the University of Sheffield's #NLProc Group is one of the UK's largest natural language processing research centres.
Joined November 2009
- Tweets 935
- Following 150
- Followers 3,000
- Likes 799
55 Photos and videos
Sheffield NLP retweeted
May 27
Excited for my invited talk at Cambridge this Friday! Thanks to @suchirsalhan & @CambridgeLTL for having me.
📊 Slides will be available after the talk: gucci-j.github.io/talks/
🗣️ ACL attendees: Join our oral session at Harbor H-I (July 5, 16:00-17:30)!
#NLProc #ACL2026
5
17
2,426
Sheffield NLP retweeted
Jun 3
New #ICML2026 paper: "On Effectiveness and Efficiency of Agentic Tool-calling and RL Training".
⚠️The methodology for measuring and improving tool-calling capabilities is fragile. We break down our analysis along two critical axes: (1) effectiveness; and (2) efficiency. 👇
1
1
13
497
Thanks @suchirsalhan for your seminar today, we loved hearing about your work on BabyLM!
x.com/suchirsalhan/status/20…
I'm really excited to be delivering a seminar at @SheffieldNLP on Tuesday about my research on Small and Human-Scale Language Modeling! Thanks @_joestacey_, @_gucciiiii, Marco Valentino and @nikaletras for the invite!
1
11
790
Sheffield NLP retweeted
Apr 28
Today, in our @SheffieldNLP seminar series, we welcome Sean MacAvaney @macavaney from @TerrierTeam at University of Glasgow to give a talk entitled "Re-thinking Re-ranking".
1
7
330
Sheffield NLP retweeted
Thrilled to see such a fantastic breakdown of our latest paper on output diversity collapse in post-training (with @Xingwei__Tan, @nikaletras, @SheffieldNLP)! 🚀
@neural_avb did a phenomenal job summarizing our work. Check it out below! 👇
Apr 23
I have been studying a lot of post-training paper to prepare for my next video.
This morning I am studying a banger from this week: "Where does output diversity collapse in post-training?"
One of the more educational papers I have read recently.
Basically after post-training LLMs often produce less varied answers than their base versions. This paper runs experiments to figure out where exactly this diversity is lost, the effects of distillation vs instruction-tuning vs RLVR, etc.
Their big hypothesis is that the diversity cannot be recovered at inference through sampling techniques or CoT. The collapse irreversibly happens during training.
3
2
5
847
Apr 22
Thanks very much to @sandropezzelle for your brilliant seminar about whether language processing in LLMs reflects human cognitive mechanisms.
2
2
12
930
Sheffield NLP retweeted
Join us on 8 June in #Sheffield to hear about #PhD students’ #AI and #ML #speech and #NLProc research!
#Keynotes from @egrefen (@GoogleDeepMind) & @wang_wenwu (@cvssp_research).
Find out more and #register (by 31 May) at slt-cdt.sheffield.ac.uk/annu…
2
1
276
Sheffield NLP retweeted
Apr 20
When we adapt LLMs to other languages using unlabeled target language data, they tend to lose their core instruction-following and reasoning abilities.
🚨 New #ACL2026 paper w/ @_gucciiiii, T. Morishita & @AlineVillav on how to mitigate this catastrophic forgetting 👇
1
3
27
3,275
Sheffield NLP retweeted
Apr 15
Our heart @SheffieldNLP will always be with *ACL, but we've also started getting serious about IR. Great to see our group's efforts in bridging #NLProc & search reflected so clearly in #SIGIR2026.
Special shout out to @wangxieric, @drmarkstevenson et al for leading this 👇
Apr 15
#SIGIR2026 accepted paper list has been released on sigir2026.org/en-AU/pages/pr…
Vibe coded the analysis again based on csrankings.org (again, by no means accurate), found that this year, Renmin University continues to be the top-1 institute, and China and US are still the top-2 countries with most publications.
3
18
1,584
Sheffield NLP retweeted
Apr 15
LLMs are great at outputting step-by-step reasoning, but is that text actually driving their conclusions, or is it just post-hoc rationalisation? How can we improve the faithfulness of LLM reasoning?
My talk at the #EACL2016 FEVER wrkshp aimed to probe a dialogue on this 🧵👇
1
1
14
1,283
Sheffield NLP retweeted
Apr 15
Excited to share my first postdoc paper with @SheffieldNLP ! 🤩
In this work we argue that supervised uncertainty quantification (UQ) needs better evaluation
Want to know more? Here's a little summary 🧵
6
9
83
7,811
Sheffield NLP retweeted
Congrats to @tylerl404 on passing his #PhD #viva on Assessing Phonetically and Phonologically Motivated Language Understanding. Thanks to supervisors Prof @chenghua_lin and Prof Rob Gaizauskas and viva examiners Dr @bjoernross (Edinburgh) and Dr @zhengyuan_nlp (Sheffield)!
1
2
6
960
We’re excited to share our latest papers accepted at @aclmeeting, @SIGIRConf, and beyond 🎉
From LLM reasoning and evaluation to multilingual NLP and fact-checking, there’s a lot we’ve been working on!
Check them out 👇
More details and paper links coming soon 🔜
1
6
28
1,785
Sheffield NLP retweeted
Happy to announce that 3 papers have been accepted to #ACL2026 Main conference! Details in the subsequent posts.
See you in San Diego!
🧷Note: I’m on the job market for Faculty/Research Fellow roles (starting April 2027)!
Details here: gucci-j.github.io/about/
(1/3) #NLProc
2
7
62
3,135
Sheffield NLP retweeted
Join us on 8 June in #Sheffield to hear about #PhD students’ #AI and #ML #speech and #NLProc research!
#Keynotes from @egrefen (@GoogleDeepMind) & @wang_wenwu (@cvssp_research).
Find out more and #register (by 31 May) at slt-cdt.sheffield.ac.uk/annu…
1
4
227
Mar 24
🎉 SheffieldNLP at @eaclmeeting
The main conference kicks off tomorrow, come say hello to our authors, we’d love to chat! 👋✨ @delvincezhang @miles_wil @nikaletras @casszzx @_joestacey_
P.S. Proud that EACL General Chair @AlineVillav is a Sheffield alum 🎓💖
#eacl2026
🎉 New paper acceptances from Sheffield NLP!
Our recent work has been accepted at @eaclmeeting , ECIR , and the CLEF Lab 🚀
Check out the poster for paper titles, authors, and links to available arXiv preprints 👇
1
11
815
Sheffield NLP retweeted
Mar 5
For our Sheffield NLP (@SheffieldNLP) seminar this week, we are delighted to have Chuan Meng (@ChuanMg) from the University of Edinburgh, sharing his research experience and his recent publication on revisiting text ranking in deep research.
1
1
11
487
Sheffield NLP retweeted
Feb 18
A day to welcome Ingo and Libo @iFromm @Bertha0103 to Sheffield @SheffieldNLP for their OMINO UK tour on “Fake Science” and information overload in Academia.
1
1
3
287
Jan 29
More #ICLR2026 good news from @SheffieldNLP 🎉
Alongside Cass’s papers, we’re excited to also have:
"Attributing Response to Context: A Jensen–Shannon Divergence Driven Mechanistic Study of Context Attribution in RAG" by @wangxieric and collaborators
arxiv.org/abs/2505.16415
Jan 26
Happy to share that two papers were accepted to #ICLR2026 🎉
A perfect cure for the Monday/January blues.
Big shout-out to my amazing collaborators 🙌
1️⃣ PerSpectra: A Scalable and Configurable Pluralist Benchmark @lucie_nlp
2️⃣ Tracing and Reversing Edits @@SeifertChristin
5
439