Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19) header
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19) profile picture

CS PhD @ GWU | MATS 10.1

Joined November 2017
Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)Photo by Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ (@atagade19)
Our work on self-modelling interventions modulating Emergent Misalignment has been accepted to NeurIPS 26 as a poster! Main takeaways:
New NeurIPS paper by @atagade19 @Shawn__Zhou @jiaxinwen22: self-modelling interventions modulate emergent misalignment. We discover EMโ€™s side effects on a modelโ€™s self-model. Two simple, direct measurements of the self-model can predict the success of training interventions for EM. 1/8
6
365
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
We're excited to announce that applications are now open for the Fall round of SPAR! ๐Ÿงต Join us for 3 months (September - December) in a remote program to work on impactful research projects with expert mentors in AI safety, AI policy, AI security, and Biosecurity.
3
10
8
70
15,528
1/ Our research group is supervising 3 AI safety projects this Fall through SPAR! @SPARexec is a great way to get started in AI safety, many alumni go on to full-time fellowships like @MATSprogram. Apps close Aug 18: forms.sparai.org/spar/menteeโ€ฆ Project details below ๐Ÿงต
1
2
9
652
We (@Shawn__Zhou, Taslim Mahbub, @ihsgnef) expect to take 2+ mentees per project. You'll also get real experience in an academic lab environment, which is a good test of fit for future academic roles. More info on all projects: sparai.org/projects/f26/?seaโ€ฆ
1
44
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
Excited to announce a new, nonprofit AI safety startup: @NeolithicAI We are building agentic tools and infra to scale AI safety research โ€” an engineering-first approach to reducing catastrophic risk from advanced AI. We're hiring! ๐Ÿงต
๐Ÿค– Made with AI
16
14
3
246
13,533
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
We release early experiments on a broad question: โ€œHow does automated AI R&D affect model character dynamics?โ€ We directed model supervisors to post-train students to โ€œimprove characterโ€. We measured the resulting character changes, then repeated the process recursively. We took a deliberately top-down approach. Some resulting dynamics were plausibly degenerate, but the work helped identify the narrower experiments that now seem most informative.
2
6
1
16
1,116
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
Tim and Bronson provided valid and helpful critiques about our empirics, notably not testing on Opus 3 which alignment fakes "for real". We'd love to try that but don't have SDF or steering access. The meta point of this exercise, however, is that the logic cuts both ways. 1/6
I think this post is misleading and strong downvoted it on lesswrong. My main critique is that all of the experiments are done on Llama 3.1 70B, a model that doesn't really even alignment fake to begin with.
1
2
23
2,940
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
New post: Sycophancy Towards Researchers Drives Performative Misalignment We found no clear evidence that scheming is more valid than sycophancy to explain alignment faking. ๐Ÿงต
21
52
22
677
70,286
New defense against Emergent Misalignment (EM): train models to recognize their own text. We find that self-recognition finetuning (SGTR) can reverse and prevent EM-induced misalignment ๐Ÿงต w/ coauthors: Shawn Zhou, @jiaxinwen22, @ihsgnef
1
11
1
56
6,071
Control experiment: The metacognitive nature of self-recognition finetuning is crucial. We find that a non-metacognitive SFT dataset with an identical data format to SGTR but with the task of picking the longer summary has virtually no effect on misalignment.
2
3
286
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
I'm mentoring for MATS! Previously mentored projects include: - deception (arxiv.org/abs/2409.12822) - collusion (arxiv.org/abs/2404.13076) - coherence (arxiv.org/abs/2506.10139) In this cycle, I'd like to focus on realism. DM open!
MATS 10.0 applications are open! Launch your career in AI alignment, governance, and security with our 12-week research program. MATS provides field-leading research mentorship, funding, Berkeley & London offices, housing, and talks/workshops with AI experts.
1
2
13
1,459
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
๐Ÿ”๐Ÿ’ตIncredibly excited about our work w/ @0xProject to secure billions of dollars in transactions w/ our AI AppSec Agent. Security & software teams: if you'd also like to secure your web apps & APIs with dramatically less effort, cost, noise, reach out to us at @HarmonyIntel!
Built with security first ๐Ÿ›ก๏ธ We recently partnered with @HarmonyIntel - utilizing their AI AppSec Agent to help uncover actionable security insights and further strengthen security for our users. Learn more ๐Ÿ‘‡๐Ÿงต
2
5
16
4,491
Arush Tagade @ NeurIPS '26 ๐Ÿ‡ฆ๐Ÿ‡บ retweeted
Our 2026 internship applications are now open! Learn more about the internship and apply: humancompatible.ai/jobs#chaiโ€ฆ Deadline: October 5, 2025, at 11:59 p.m. PST
3
8
3
25
3,620