Joined November 2012
- Tweets 68
- Following 589
- Followers 145
- Likes 388
6 Photos and videos
Pinned Tweet
4 Jun 2025
LLMs adjust their responses to their given persona and to their interlocutor’s…
…but what happens when they’re from different planets? 🧔🦸♂️
Find out in our #ACL2025NLP (main) paper!
👉Preprint: arxiv.org/abs/2505.24613
What happens when Harry meets Superman instead of Sally for a chat? 🧔🦸♂️
We explore how LLMs adapt their responses not just to their given persona, but also to their interlocutors’.
Excited to share our #ACL2025NLP (main) paper by @docchipinti8 @m_guerini @MalvinaNissim
1/🧵
3
214
Daniela Occhipinti retweeted
Jan 13
🎶One more “song”?
Come see our poster next week at #AAAI2026 on our paper:
📣 “Don’t Stop the Multi-Party! On Generating Synthetic Written Multi-Party Conversations with Constraints”
With @m_guerini, @brulepri, @gg42554, @sara_hlt
🧵
arxiv.org/abs/2502.13592
1
8
16
901
Daniela Occhipinti retweeted
8 Oct 2025
We found a troubling emergent behavior in LLM.
💬When LLMs compete for social media likes, they start making things up
🗳️When they compete for votes, they turn inflammatory/populist
When optimized for audiences, LLMs inadvertently become misaligned—we call this Moloch’s Bargain
845
1,945
9,648
1,340,248
Daniela Occhipinti retweeted
4 Aug 2025
We're back from #ACL2025NLP and still buzzing from how well CSHAM went! Huge thanks to everyone who attended, asked questions, and made it such a great tutorial!
6
11
269
📢 Do you want to join our group? We opened **two** calls for a fully funded PhD. Details are in the image and at the following links.
Calls: iecs.unitn.it/education/admi…
PhD Details: iecs.unitn.it/education/admi…
Deadline: August 22nd, 2025, hrs. 04:00 PM (CEST)
9
5
1,691
28 Jul 2025
It was amazing meeting so many people this morning and chatting about how LLMs adapt not just to their given persona, but to the interlocutor’s too. I really enjoyed the poster session and loved everyone’s insights! 🚀
For more details: aclanthology.org/2025.acl-lo…
#ACL2025NLP
1
12
287
30 May 2025
Beyond happy for this result! 🥳🚀
🎉 The LanD group is excited to share that 3 of our papers made it into #ACL2025NLP!
Massive congrats to the authors: @docchipinti8, @russ_eau, @stefano_menini, @_FaribaSadeghi_, @ziorufus, @m_guerini, @MalvinaNissim.
More details and links to the papers coming soon! 👀
2
1
5
149
Daniela Occhipinti retweeted
6 Dec 2024
4
22
1,787
Daniela Occhipinti retweeted
5 Dec 2024
3
5
373
Today we will present three of our recent works at #CLiCit2024 with @helena_bonaldi, @docchipinti8, @russ_eau.
Don't miss the poster session today at 10 👀
3
6
239
To click it, or not to click it: that is the question.
Our #CLiC-it2024 contribution is finally out: "To Click It or Not to Click It: An Italian Dataset for Neutralising Clickbait Headlines" authored by @russ_eau , @oaraquei , and @m_guerini .
1
3
8
1,134
Daniela Occhipinti retweeted
10 Oct 2024
💭Remember the Stanford Prison Experiment?
📣“I Want to Break Free! Anti-Social Behavior and Persuasion Ability of LLMs in Multi-Agent Settings with Social Hierarchy”, our new pre-print, is out!
W/@penzo_nicolo, @itakello, @robdessi, @m_guerini, @brulepri, @stjaco
Check this🧵
3
7
37
5,151
Daniela Occhipinti retweeted
7 Oct 2024
🚨 Our paper “Is Safer Better? The Impact of Guardrails on the Argumentative Strength of LLMs in Hate Speech Countering” has been accepted at #EMNLP2024!
With @grexit_ @NBenjaminOcampo @ECabrio @serena_villata @m_guerini
arxiv.org/abs/2410.03466
A thread 🧵
2
8
37
3,346
Daniela Occhipinti retweeted
1 Oct 2024
🍸Haven't you had enough rounds, Llama?
📣"Do LLMs suffer from Multi-Party Hangover? A Diagnostic Approach to Addressee Recognition and Response Selection in Conversations", our #EMNLP2024 paper!
With @mariamsajedinia, @brulepri, @sara_hlt, @m_guerini
🧵
arxiv.org/abs/2409.18602
3
8
27
2,445
14 Aug 2024
Dear Reviewer 2, these are NOT valid reasons to reject o poorly rate a paper! #ACL2024NLP
1
3
7
1,070
Daniela Occhipinti retweeted
remember GloVe embeddings? of course you do, they are still being used (and we teach them for their simplicity and elegance), as @chrmanning pointed out in his speech. Congratulations on the 10yr Test-of-Time award! #ACL2024NLP
14 Aug 2024
The GloVe paper wins the 10yr test of time award. #ACL2024NLP
2
3
46
6,020
13 Aug 2024
I am happy to have presented our work on the impact of human post-editing of synthetic data on dialogical language models at #ACL2024NLP! Enjoyed discussing our findings with many interested people!
Want to learn more? Check out our paper!
👉 aclanthology.org/2024.findin…
1
22
1,108
12 Aug 2024
If you are at #ACL2024NLP, come say hi! I will be presenting our poster "Fine-tuning with HED-IT: The impact of human post-editing for dialogical language models" at Findings In-Person Poster Session 3, tomorrow at 12:15! 📝 #ACL2024
2
15
962