@a-krishnan.bsky.social
π€ 44
π₯ 68
π 8
Master student at Saarland university
Join me and
@mariusmosbach.bsky.social
to chat about our work on frequency effects in unlearning β and how
@ai2.bsky.social
's Olmo helped us gain key insights. π¬ AMA: Tue, Oct 28 β 8:00 PT / 16:00 CEST π‘ Bring your questions! π
discord.gg/ai2
add a skeleton here at some point
about 2 months ago
0
2
0
We're presenting βNot all data are unlearned equallyβ at
#COLM2025
! We show that data properties shape how LLMs forget β stop by to chat more! π Wednesday, Oct 8 π 4:30β6:30 pm π poster #710 (session 4) paper:
arxiv.org/abs/2504.05058
Work with
@mariusmosbach.bsky.social
@sivareddyg.bsky.social
add a skeleton here at some point
2 months ago
0
0
0
reposted by
Benno Krojer
4 months ago
very happy to see the trend of a Behind the Scenes section catching on! transparent & honest science π love the detailed montreal spots mentioned consider including such a section in your next appendix! (paper by
@a-krishnan.bsky.social
arxiv.org/pdf/2504.050...
)
1
7
2
reposted by
Gaurav Kamath
5 months ago
Our new paper in
#PNAS
(
bit.ly/4fcWfma
) presents a surprising findingβwhen words change meaning, older speakers rapidly adopt the new usage; inter-generational differences are often minor. w/ Michelle Yang, βͺ@sivareddyg.bsky.socialβ¬ ,
@msonderegger.bsky.social
β¬ and
@dallascard.bsky.social
β¬π(1/12)
3
34
19
π’
#SpeechTech
&
#SpeechScience
researchers! We are thrilled to announce that Prof. Karen Livescu will keynote our Special Session on Interpretable Audio and Speech Models at
#Interspeech2025
: "What can interpretability do for us (and what can it not)?" ποΈ Aug 18, 11:00
@interspeech.bsky.social
loading . . .
Announcements
Keynote Speaker Announcement π 30.07.2025We are delighted to announce the keynote speech t`hat will happen at the special session!Speaker: Prof. Karen Livescu, Toyota Technological Institute at Ch...
https://sites.google.com/view/interspeech2025-interpret/announcements
5 months ago
0
3
2
reposted by
Gaofei Shen
7 months ago
I am excited to announce that my paper "On the reliability of feature attribution methods for speech classification" has been accepted to
#Interspeech2025
! Co-authors:
@hmohebbi.bsky.social
, Arianna Bisazza, Afra Alishahi,
@grzegorz.chrupala.me
Find the preprint here:
arxiv.org/abs/2505.16406
loading . . .
On the reliability of feature attribution methods for speech classification
As the capabilities of large-scale pre-trained models evolve, understanding the determinants of their outputs becomes more important. Feature attribution aims to reveal which parts of the input elemen...
https://arxiv.org/abs/2505.16406
1
10
3
reposted by
Vagrant Gautam
7 months ago
Come to my keynote tomorrow at the first official
@queerinai.com
workshop at
#NAACL2025
to hear about how trans languaging is complex and cool, and how this makes it extra difficult to process computationally. I will have SO many juicy examples!
3
44
14
reposted by
Michael Hahn
7 months ago
Chain-of-Thought (CoT) reasoning lets LLMs solve complex tasks, but long CoTs are expensive. How short can they be while still working? Our new ICML paper tackles this foundational question.
2
12
2
reposted by
Benno Krojer
8 months ago
A must-read for anyone in NLP right now
add a skeleton here at some point
1
6
1
reposted by
Mila - Institut quΓ©bΓ©cois d'IA
8 months ago
Congratulations to Mila members
@adadtur.bsky.social
, Gaurav Kamath and
@sivareddyg.bsky.social
for their SAC award at NAACL! Check out Ada's talk in Session I: Oral/Poster 6. Paper:
arxiv.org/abs/2502.05670
loading . . .
0
13
10
reposted by
Siva Reddy
8 months ago
Incredibly proud of my students
@adadtur.bsky.social
and Gaurav Kamath for winning a SAC award at
#NAACL2025
for their work on assessing how LLMs model constituent shifts.
add a skeleton here at some point
1
17
5
reposted by
Yanai Elazar
8 months ago
π‘ New ICLR paper! π‘ "On Linear Representations and Pretraining Data Frequency in Language Models": We provide an explanation for when & why linear representations form in large (or small) language models. Led by
@jackmerullo.bsky.social
, w/
@nlpnoah.bsky.social
&
@sarah-nlp.bsky.social
3
42
15
reposted by
Xing Han Lu
8 months ago
DeepSeek-R1 Thoughtology: Letβs <think> about LLM reasoning 142-page report diving into the reasoning chains of R1. It spans 9 unique axes: safety, world modeling, faithfulness, long context, etc. Now on arxiv:
arxiv.org/abs/2504.07128
1
6
1
reposted by
Xing Han Lu
8 months ago
AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories We are releasing the first benchmark to evaluate how well automatic evaluators, such as LLM judges, can evaluate web agent trajectories.
1
7
5
reposted by
Marius Mosbach
8 months ago
Checkout Benno's notes about our impact of interpretability paper π. Also, we are organizing a workshop at
#ICML2025
which is inspired by some of the questions discussed in the paper:
actionable-interpretability.github.io
add a skeleton here at some point
0
11
3
reposted by
Marius Mosbach
8 months ago
Check out our new paper on unlearning for LLMs π€. We show that *not all data are unlearned equally* and argue that future work on LLM unlearning should take properties of the data to be unlearned into account. This work was lead by my intern
@a-krishnan.bsky.social
π:
arxiv.org/abs/2504.05058
1
33
7
reposted by
VLMs4All - CVPR 2025 Workshop
9 months ago
π’Excited to announce our upcoming workshop - Vision Language Models For All: Building Geo-Diverse and Culturally Aware Vision-Language Models (VLMs-4-All) @CVPR 2025! π
sites.google.com/view/vlms4all
1
17
15
reposted by
Xing Han Lu
9 months ago
Agents like OpenAI Operator can solve complex computer tasks, but what happens when users use them to cause harm, e.g. spread misinformation? To find out, we introduce SafeArena (
safearena.github.io
), a benchmark to assess the capabilities of web agents to complete harmful web tasks. A thread π
1
17
12
π’
#SpeechTech
&
#SpeechScience
researchers! β³ Reminder: The
#Interspeech2025
deadline is approaching! π If your work focuses on interpretability in speech & audio, submit through our Special Session and showcase your research! π€
#Interpretability
@interspeech.bsky.social
loading . . .
Home
Introduction Audio and speech technology has recently achieved unprecedented success in real-world applications, driven primarily by self-supervised pre-training of large neural networks on massive da...
https://sites.google.com/view/interspeech2025-interpret/home
11 months ago
1
1
0
you reached the end!!
feeds!
log in