Sign in

Chantal

@chantalsh.bsky.social
64 followers 60 following 17 posts

PhD (in progress) @ Northeastern! NLP 🤝 LLMs she/her

PostsRepliesMedia
Chantal @chantalsh.bsky.social · 24/10/2025
(4/n) More info here! Read our paper: arxiv.org/abs/2509.21155 Paper site: cshaib.github.io/syntax_domai... Thank you to all my wonderful co-authors; happy to continue chatting about any of this!
arxiv.org
Learning the Wrong Lessons: Syntactic-Domain Spurious Correlations in Language Models
For an LLM to correctly respond to an instruction it must understand both the semantics and the domain (i.e., subject area) of a given task-instruction pair. However, syntax can also convey implicit i...
000
Chantal @chantalsh.bsky.social · 24/10/2025
(3/n) Perhaps more strikingly, unintended syntactic-domain correlations can be exploited to bypass model refusals (e.g., OLMo-2-Instruct 7B here)
100
Chantal @chantalsh.bsky.social · 24/10/2025
(2/n) This has important implications for model generalization and safety! We show that this occurs in instruction-tuned models, and propose an evaluation to test for this type of brittleness.
100
Chantal @chantalsh.bsky.social · 24/10/2025
(1/n) Models learn to rely on *syntactic templates* (frequent patterns of POS tags) that co-occur with particular domains. LLMs can inadvertently learn "If I see this syntactic pattern it’s domain X" rather than "If I see this semantic content, do task Y."
100
Chantal @chantalsh.bsky.social · 24/10/2025
Syntax that spuriously correlates with safe domains can jailbreak LLMs - e.g. below with GPT4o mini Our paper (co w/ Vinith Suriyakumar) on syntax-domain spurious correlations will appear at #NeurIPS2025 as a ✨spotlight! + @marzyehghassemi.bsky.social, @byron.bsky.social, Levent Sagun
363
Chantal @chantalsh.bsky.social · 24/09/2025
(7/7) For more details, please check out our pre-print!
040
Chantal @chantalsh.bsky.social · 24/09/2025
(6/7) LLMs are terrible at detecting their own slop: GPT-5, Deepseek-V3, and o3-mini rarely assign a label of "slop" (avg. 6% of documents), whereas humans marked 34% of texts as "slop."
150
Chantal @chantalsh.bsky.social · 24/09/2025
(5/7) We lack good/reliable automatic text metrics for 3 of the 5 most important slop features: relevance, coherence, and tone. :-(
110
Chantal @chantalsh.bsky.social · 24/09/2025
(4/7) Different domains have different slop signatures. In news articles, coherence, density, relevance, and tone issues predict slop. In Q&A tasks, it's factuality and structure. Context matters!
120
Chantal @chantalsh.bsky.social · 24/09/2025
(3/7) Humans can spot "sloppy text", but may have differing thresholds on overall assessments. But our annotators consistently flagged the same problematic passages, suggesting we know it when we see it...
110
Chantal @chantalsh.bsky.social · 24/09/2025
(2/7) TL;DR: Measuring the construct of slop is difficult! While somewhat subjective and domain-dependent, it boils down to three key factors: information quality, density, and stylistic choices. We introduce a taxonomy for slop.
120
Chantal @chantalsh.bsky.social · 24/09/2025
"AI slop" seems to be everywhere, but what exactly makes text feel like "slop"? In our new work (w/ @tuhinchakr.bsky.social, Diego Garcia-Olano, @byron.bsky.social ) we provide a systematic attempt at measuring AI "slop" in text! arxiv.org/abs/2509.19163 🧵 (1/7)
13317
Chantal @chantalsh.bsky.social · 24/09/2025
(5/7) We lack good/reliable automatic text metrics for 3 of the 5 most important slop features: relevance, coherence, and tone. :-(
000
Chantal @chantalsh.bsky.social · 24/09/2025
(4/7) Different domains have different slop signatures. In news articles, coherence, density, relevance, and tone issues predict slop. In Q&A tasks, it's factuality and structure. Context matters!
100
Chantal @chantalsh.bsky.social · 24/09/2025
(3/7) Humans can spot "sloppy text", but may have differing thresholds on overall assessments. But our annotators consistently flagged the same problematic passages, suggesting we know it when we see it...
100
Chantal @chantalsh.bsky.social · 24/09/2025
(2/7) TL;DR: Measuring the construct of slop is difficult! While somewhat subjective and domain-dependent, it boils down to three key factors: information quality, density, and stylistic choices. We introduce a taxonomy for slop.
100
Chantal @chantalsh.bsky.social · 10/03/2025
I'm searching for some comp/ling experts to provide a precise definition of “slop” as it refers to text (see: corp.oup.com/word-of-the-...) I put together a google form that should take no longer than 10 minutes to complete: forms.gle/oWxsCScW3dJU... If you can help, I'd appreciate your input! 🙏
corp.oup.com
Oxford Word of the Year 2024 - Oxford University Press
The Oxford Word of the Year 2024 is 'brain rot'. Discover more about the winner, our shortlist, and 20 years of words that reflect the world.
0108
Reposted by Chantal
Somin W @sominw.bsky.social · 11/02/2025
📢 Can we trace a small distilled model back to its teacher? 🤔New work (w/ @chantalsh.bsky.social, @silvioamir.bsky.social & @byron.bsky.social) finds some footprints left by LLMs in distillation! [1/6] 🔗 Full paper: arxiv.org/abs/2502.06659
arxiv.org
Who Taught You That? Tracing Teachers in Model Distillation
Model distillation -- using outputs from a large teacher model to teach a small student model -- is a practical means of creating efficient models for a particular task. We ask: Can we identify a stud...
182