Research Scientist - interpretability/alignment @MetaAI MSL/GenAI Trust/Responsible AI. PhD @UTAustin 22. @IBMResearch 20, @GoogleAI 18/19, @datascifellows 16

Joined July 2009
21 Photos and videos
Pinned Post
Excited to announce we'll be hosting the 2nd U&Me workshop at #ICCV2025 2025 in October. If you work in unlearning or model editing you should consider submitting something or participating in the challenge ( top winners will be co-authors on workshop challenge paper )
๐Ÿ”ฅ News - 2nd Unlearning and Model Editing Workshop and Challenge at #ICCV2025 ๐Ÿ“ƒ Call for papers ready and OpenReview accepting submissions: bit.ly/4knWGv2 ๐Ÿงฉ New challenge on #Unlearning bit.ly/43GkzbK / unlearning.iab-rubric.org/ Best performers in paper!
3
11
1,974
I second Iacopo's sentiment on being a co-organizer of this workshop on Unlearning and Model Editing . Paper submissions due July 9 for folks working in this important space! #mechanisticinterpretability #unlearning #ECCV2026
๐Ÿง  Proud to be co-organizing U&ME again โ€” now in its 3rd edition! Unlearning & model editing have gone from niche to essential: privacy, safety, keeping models current. If that's your space, come share it in Malmรถ ๐Ÿ‡ธ๐Ÿ‡ช ๐Ÿ‘‡ #ECCV #computervision #unlearning
2
545
Diego Garcia-Olano reposted
How did #ECCV reviews treat you? ๐Ÿ˜ฌ If you're not that happy โ€” and you work on #unlearning, #modelediting, #modelmerging, or #interpretability โ€” consider submitting to the 3rd Workshop and Challenge on Unlearning and Model Editing at @eccvconf We accept both full papers and extended abstracts, so there's a format for every contribution. ๐Ÿ“… Submission deadline: 9 July 2026 (AoE) ๐Ÿ”—openreview.net/group?id=thecโ€ฆ
๐Ÿ“ขCPF: U&ME Workshop @ ECCV 2026 @eccvconf We invite submissions on machine unlearning, model editing, and related topics including efficient adaptation, and responsible AI. Details at: sites.google.com/view/u-and-โ€ฆ @_iAc #ECCV #ECCV2026 #Unlearning #AIsafety #ResponsibleAI
3
3
1,265
Diego Garcia-Olano reposted
๐Ÿง  Excited to announce ๐•Œ&๐•„๐”ผ ๐Ÿš๐Ÿ˜๐Ÿš๐Ÿžโ€” ๐˜๐—ต๐—ฒ ๐Ÿฏ๐—ฟ๐—ฑ ๐—ช๐—ผ๐—ฟ๐—ธ๐˜€๐—ต๐—ผ๐—ฝ & ๐—–๐—ต๐—ฎ๐—น๐—น๐—ฒ๐—ป๐—ด๐—ฒ ๐—ผ๐—ป ๐—จ๐—ป๐—น๐—ฒ๐—ฎ๐—ฟ๐—ป๐—ถ๐—ป๐—ด ๐—ฎ๐—ป๐—ฑ ๐— ๐—ผ๐—ฑ๐—ฒ๐—น ๐—˜๐—ฑ๐—ถ๐˜๐—ถ๐—ป๐—ด @eccvconf Organized with @Hussain68018934 @THassner @thakralkartik78 @BrigliaRosaria @MayankVatsa3 @dgolano @sijialiu17 ๐Ÿ“… September 8โ€“9, 2026 ๐Ÿ“ Malmรถ, Sweden How do we fix, update & align large generative models โ€” without retraining from scratch? ๐ŸŽค Confirmed Speakers: Yezhou Yang โ€” Arizona State University Fabio Galasso โ€” Sapienza University of Rome William Shen โ€” University of Cambridge More speakers coming soon! Stay tuned for submission details ๐Ÿ‘‡ sites.google.com/view/u-and-โ€ฆ #ECCV2026 #MachineLearning #ModelEditing #MachineUnlearning #GenerativeAI #DeepLearning
7
11
1,764
Check out findings from our new paper on LLM knowledge distillation memorization dynamics! Great job @JaydeepBorkar !
Excited to share my work at Meta. Knowledge Distillation has been gaining traction for LLM utility. We find that distilled models don't just improve performance, they also memorize significantly less training data than standard fine-tuning (reducing memorization by >50%). ๐Ÿงต
3
10
1,098
Grateful to our co-authors @xinchiqiu, @jade_lei_yu , Yuchen Zhang, Oliver Aobo Yang, @Nunushik, @nicola_cancedda. Come say hi if you're around at the workshop on Sunday!
1
50
CASAL is an interpretability-inspired methodย for improving alignment & mitigating hallucinations of multi-modal, dense & MOE based LLMs that is both effective & highly parameter and sample efficient. More lightweight than LORA SFT & doesn't rely on inference time steering ...
1
54
Heading to #NeurIPS2025 in San Diego this weekend to discuss my intern @winnieyangwn & our team's work accepted at the #MechInterp Workshop. Hallucination Reduction with CASAL: Contrastive Activation Steering for Amortized Learning ( arxiv.org/abs/2510.02324 ) ...
1
4
203
If you're at @ICCVConference come join us and say hello tomorrow afternoon Oct 19, from 1-6pm for the 2nd Unlearning and Model Editing workshop. Lots of great keynotes and papers being presented. sites.google.com/view/u-and-โ€ฆ
1
231
Diego Garcia-Olano reposted
๐ŸŒบ Join us at the U&ME workshop at @ICCVConference in Hawaii on Oct 19, 2025, 1-6 pm. ๐ŸŽ™ Yair Adato โ€” CEO, BRIA AI ๐ŸŽ™ Sijia Liu โ€” Red Cedar Distinguished Associate Prof, Michigan State University ๐Ÿ”— sites.google.com/view/u-and-โ€ฆ @THassner @dgolano @MayankVatsa3 @thakralkartik78
4
12
2,036
Diego Garcia-Olano reposted
"AI slop" seems to be everywhere, but what exactly makes text feel like slop? In our new work (w/ @TuhinChakr, @dgolano, @byron_c_wallace) we provide a systematic attempt at measuring AI slop in text! arxiv.org/abs/2509.19163 ๐Ÿงต (1/7)
help me fix get-4o slop reply with examples of slop behavior just a single sentence nothing crazy what annoys you what makes you wanna frisbee your laptop into a river i'll respond to every comment rt so we can maximize slop feedback help me de-sloptimize our models go
14
37
220
34,776
We are excited for this year's workshop and competition. More info at sites.google.com/view/u-and-โ€ฆ
1
1
114
There is a week left for submissions to the 2nd edition of the Unlearning & Model Editing (U&Me) workshop at #ICCV2025. We are accepting both long (archival) & short(non-archival) submissions. Last year's U&Me workshop with @davidbau & @rita_cucchiara at #ECCV was a joy & ...
1
2
2
547
Well said. Arbitrarily cutting off Chinese or any international students from coming to the US is cruel and runs counter to what has helped enrich the country financially and culturally for the past nearly 100 years.
Revoking visas to Chinese PhD students is economically shortsighted and inhumane. Most Chinese PhD students stay in the U.S. after graduation (first image, stats from 2022). They're staying and building technology in the U.S., not taking it to China. Immigrant students create startup companies that employ Americans (second image, stats from 2018). I couldn't find stats for Chinese students specifically but anecdotally they are a significant force powering GenAI and LLM companies. These are the next $1B companies that are going to hire thousands of people even as other jobs are automated. There are no real "secrets" in academic GenAI. Everything is open-source. What is the security risk? It's a much higher risk to give DeepSeek and others an advantage in hiring the most talented young researchers in these areas by preventing them from coming to the U.S. Revoking visas for these students destroys American innovation for no benefit. It's also unfair and inhumane. In my experience, Chinese PhD students are hardworking young researchers interested in their fields of study and being part of the cutting edge of academic and industrial innovation. They aren't geopolitical pawns. Stop implementing policies that treat them that way.
1
298
Diego Garcia-Olano reposted
2nd edition of U&ME workshop will be atย #ICCV2025. After the success of 1st workshop atย #ECCV2024, weโ€™re gearing up for an even more exciting programโ€”including our first-ever competitionย ... details coming soon With @_iAc @dgolano @THassner @thakralkartik78 #Unlearning #genAI
4
13
658
Diego Garcia-Olano reposted
Why is interpretability the key to dominance in AI? Not winning the scaling race, or banning China. Our answer to OSTP/NSF, w/ Goodfire's @banburismus_ Transluce's @cogconfluence MIT's @dhadfieldmenell resilience.baulab.info/docs/โ€ฆ Here's why:๐Ÿงต โ†˜๏ธ
1
69
309
37,247
Some pre-reqs: Students must be enrolled in a PhD program, returning at the end of the internship and have published on the topics. There are other RS internships on the Meta careers site and I've reposted a few.
4
812
Diego Garcia-Olano reposted
I am looking for a Research Scientist intern for 2025. If you have already published work that involves understanding behaviours of AI models looking at their parameters and activations, I would like to hear from you. metacareers.com/jobs/5560633โ€ฆ
5
48
319
45,109
Diego Garcia-Olano reposted
I am looking for a summer or fall research intern next year! If you're excited about attacking data / evaluation questions at scale, please reach out to me with a short blurp and your CV! Both email and DM are good, have some patience for the response pls :).
4
19
264
38,329
Hi all, I'll be hosting a Research Scientist intern for 2025 with a focus on exploring LLM safety alignment and understanding potentially using explainability methods. If that sounds of interest apply/reach out! metacareers.com/jobs/9010694โ€ฆ
7
35
205
27,885