Lisa Dunlap @lisabdunlap Twitter profile

Last Seen Profiles

@DirtyRobotWorks

@yun__trihiki

@RealBenFulford

@hyckvamp

@CarrinMarion

@GraceS8714

@HawcLi

@Jekkah

@CPSharksMBB

@imma_PC7

@rolandmangold

@ktovsak

@bokeplokalmalam

@khldds

@JoeOjewumi100

@kaname__r

@Megasukoco69

@Otrius_

@MartinKove

@DTaniguchiAstro

@praliphadaa

@Timothy97158876

@Ameena_26B

@Arian_sabz

@VmaniakJ

@MovrNews

@Icaakkkk_a

@NadiaStaikos

@kuwa_d2z

@V4guely

@D0llywood

@pedropi80224285

@AnaBeat4797

@KhafurSali93786

@VE2Korg

@muzamil903379

Lisa Dunlap

@lisabdunlap

9 months

Highly recommend you ask people to sign your poster tube at conferences. I tried it for the first time this NeurIPS and it feels like I’m that cool kid who had a cast signed by the whole class :D

3

12

176

Lisa Dunlap

@lisabdunlap

9 months

I finally feel like an ML researcher.

AK

@_akhaliq

9 months

Describing Differences in Image Sets with Natural Language paper page: How do two sets of images differ? Discerning set-level differences is crucial for understanding model behaviors and analyzing datasets, yet manually sifting through thousands of

1

34

155

5

3

143

Lisa Dunlap

@lisabdunlap

9 months

[1/5] Introducing VisDiff - an #AI tool that describes differences in image sets with natural language. VisDiff can summarize model failures, compare models, find nuanced dataset differences, discover what makes an image memorable, and so much more!

4

30

126

Lisa Dunlap

@lisabdunlap

11 months

I feel like the academic equivalent of the iPhone alarm noise is the slack message “found this recent work on arxiv, seems similar to what you have been working on”

1

4

54

Lisa Dunlap

@lisabdunlap

11 months

My advisor was surprised that my code has docstrings and i realized (1) my advisor has low standards and (2) i have a ton of docstrings because i need to explain what i'm doing so copilot can code it for me

5

0

47

Lisa Dunlap

@lisabdunlap

9 months

At NeurIPS? Stop by my poster this afternoon (5:15-7:15)! #437 : Diversify Your Vision Datasets with Automatic Diffusion-based Augmentation

Diversify Your Vision Datasets with Automatic Diffusion-Based Augmentation

Many fine-grained classification tasks, like rare animal identification, have limited training data and consequently classifiers trained on these datasets often fail to generalize to variations in...

arxiv.org

1

46

Lisa Dunlap

@lisabdunlap

1 year

A fantastic lecture from @jefrankle taught me that (yet again) random sampling is the strongest baseline out there. I’m getting a PhD at one of the best universities in the world, and I’m going to be bested by np.choice

2

39

Lisa Dunlap

@lisabdunlap

10 months

Had a great BAIR salon on Science and AI! I will say there is nothing quite like 75 scientists (including students and faculty from CS and bio, a dean, and a Nobel laureate, and many more) in a room debating whether GPT will solve science :P

0

4

33

Lisa Dunlap

@lisabdunlap

3 months

We added images to chatbot arena! Here’s another fun demo:

lmsys.org

@lmsysorg

3 months

Exciting news - Chatbot Arena now supports image uploads📸 Challenge GPT-4o, Gemini, Claude, and LLaVA with your toughest questions. Plot to code, VQA, story telling, you name it. Let's get creative and have fun! Leaderboard coming soon. Credits to builders @chrischou03

13

67

421

0

9

34

Lisa Dunlap

@lisabdunlap

4 months

Battle with the mysterious gpt2! I even added a comically large vote button to the huggingface leaderboard so people know what to click :P

Chat with Open Large Language Models

lmarena.ai

Ethan Mollick

@emollick

4 months

There is a mysterious new model called gpt2-chatbot accessible from a major LLM benchmarking site. No one knows who made it or what it is, but I have been playing with it a little and it appears to be in the same rough ability level as GPT-4. A mysterious GPT-4 class model? Neat!

95

212

1K

4

7

29

Lisa Dunlap

@lisabdunlap

4 months

After a grueling few days of having to click accept to view the chatbot leaderboard, we put it back on HF :p

LMSys Chatbot Arena Leaderboard - a Hugging Face Space by lmsys

huggingface.co

0

5

28

Lisa Dunlap

@lisabdunlap

5 months

In the first year of my PhD I was obsessed with this idea of live benchmarks which would apply filtering to a stream of data to create eval sets which always represent the current world. I completely failed at this ofc, but happy to see a much better version come out :)

lmsys.org

@lmsysorg

5 months

Introducing Arena-Hard – a pipeline to build our next generation benchmarks with live Arena data. Highlights: - Significantly better separability than MT-bench (22.6% -> 87.4%) - Highest agreement to Chatbot Arena ranking (89.1%) - Fast & cheap to run ($25) - Frequent update

20

122

636

0

2

26

Lisa Dunlap

@lisabdunlap

11 months

Me: "Why don't you caption this image?" LLaVA 1.5: "I'm sorry, I cannot caption this image as it is a close-up of a bear climbing a tree, and it is not possible to provide a caption that accurately describes the scene without potentially causing harm to the bear or the tree."

1

6

22

Lisa Dunlap

@lisabdunlap

5 months

Excited to be a part of creating the category based leaderboard! Now I have a ton of category-based model insights and Gradio knowledge

lmsys.org

@lmsysorg

5 months

We tag all the conversations containing code snippets in Coding Arena. In this domain, we find GPT-4-Turbo performs even stronger. This aligns with the recent finding in challenging coding benchmark such as LiveCodeBench by You can also easily view

3

10

115

2

4

22

Lisa Dunlap

@lisabdunlap

1 year

My findings from an afternoon of playing around with captioning/VQA models - (1) they don't like captioning (2) they have the attitude of an edgy teen If you want to laugh at the audacity of BLIP/InstructBLIP, check out these examples:

1

3

18

Lisa Dunlap

@lisabdunlap

10 months

I would like the throw in a unicorn we asked GPT-4 to make right after it’s release that has been hanging in our lab ever since

Dimitris Papailiopoulos

@DimitrisPapail

10 months

I'm sure you've wondered: Can GPT-4v draw a TikZ Unicorn if we give it visual feedback? I am here to settle this open problem An 8-part 🧵on my attempt to get GPT-4v to draw a🦄 as good as @SebastienBubeck et al's, when given multiple rounds for improvement. TL;DR: i failed

5

16

142

2

1

17

Lisa Dunlap

@lisabdunlap

4 months

Can confirm, very good.

Sam Altman

@sama

4 months

it is a very good model (we had a little fun with the name while testing)

51

184

2K

2

1

14

Lisa Dunlap

@lisabdunlap

4 months

, please stop reminding me that there is another L. Dunlap who is doing far cooler research

0

13

Lisa Dunlap

@lisabdunlap

6 months

So, 1. My friend @conor_power23 and I have a podcast 2. I have Tourette's and we are still figuring out the proper accommodations, so there is some sniffling n' things in the recordings 3. We got separate mics for everyone so the audio should improve

Shreya Shankar

@sh_reya

6 months

My friends @lisabdunlap and @conor_power23 run a podcast called “Thinking About Thinking About Computers,” chatting with CS PhD students about their research journey & life outside of work. Great resource (especially if you’re considering a PhD program):

2

10

79

1

12

Lisa Dunlap

@lisabdunlap

5 months

We need to get some good matching outfits for our talk; suggestions are very welcome

Yuhui Zhang

@Zhang_Yu_hui

5 months

Super excited to share that VisDiff has been accepted to #CVPR2024 and selected as an oral (90/11,532)! We will give a 15-min presentation going through the methods and exciting applications enabled by VisDiff. See you in Seattle!

1

13

57

1

12

Lisa Dunlap

@lisabdunlap

3 months

I am going to be in CVPR next week, let me know if you want to meet up!

0

11

Lisa Dunlap

@lisabdunlap

7 months

@deliprao @profjoeyg @beenwrekt One of my favorite group meeting presentations: telling Joey I was going to explain diffusion models then spent an hour convincing everyone we should get TikTok famous (slide for evidence)

4

2

11

Lisa Dunlap

@lisabdunlap

2 months

Blog:

The Multimodal Arena is Here! | LMSYS Org

lmsys.org

@lmsysorg

2 months

🔥Exciting News — we are thrilled to announce Chatbot Arena’s Vision Leaderboard! Over the past 2 weeks, we’ve collected 17K+ votes across diverse use cases. Highlights: - GPT-4o leads the way, followed by Claude 3.5 Sonnet in #2 and Gemini 1.5 Pro in #3 - Open model

15

77

447

0

2

11

Lisa Dunlap

@lisabdunlap

6 months

Check out our and recent podcast episode with the amazing @shmu_h talking about Art & CS, AI & copyright, and the pros and cons of having an emotional connection with your research. Youtube: Spotify:

#13 - Shm Almeda: Art & Computer Science, AI & Copyright, NFTs, and HCI by Thinking About Computers

A conversation with Shm Garanganao Almeda, a 4th year PhD student at UC Berkeley studying the intersection of art and human-computer interaction. We discuss training generative AI without artist's...

spotifyanchor-web.app.link

1

10

Lisa Dunlap

@lisabdunlap

11 months

My new favorite hobby: helping improve people’s plots and figures. Fun to do and the payoff is so satisfying.

4

0

10

Lisa Dunlap

@lisabdunlap

9 months

Time for the obligatory “Im going to NeurIPS” tweet. Please reach if your interested in meeting up!

0

10

Lisa Dunlap

@lisabdunlap

3 months

Can’t wait for the AP Test on prompt engineering to drop

0

9

Lisa Dunlap

@lisabdunlap

4 months

It’s also just fun to play around with cool models… if I can learn interesting differences between models people use everyday I’m happy to be a horribly paid data scientist

Wei-Lin Chiang

@infwinston

4 months

A bit disappointed to see the bigger picture missed. The goal of the Arena project is to build a better evaluation platform to advance the field. We offer free LLM services, leaderboard insights, and open conversation & feedback data for better evals. Our aim is to bring value

12

3

51

1

9

Lisa Dunlap

@lisabdunlap

11 months

Spooky season @berkeley_ai

1

0

9

Lisa Dunlap

@lisabdunlap

7 months

How TA'ing computer vision is going: tried helping a student in office hours only for him tell me I was wrong because "that's just how images work"

2

0

8

Lisa Dunlap

@lisabdunlap

11 months

Sometimes I feel like the LLM academic landscape is just a ton of CS people discovering psychology exists

0

1

7

Lisa Dunlap

@lisabdunlap

6 months

Congrats @tianjun_zhang on some amazing work! Fun fact: Joeys group made Tianjun give a practice talk for this work while wine tasting a few months back. The waiter was quite confused…

Tianjun Zhang

@tianjun_zhang

6 months

📢 Excited to release RAFT: Retriever-Aware FineTuning for domain-specific RAG, a three-way collaboration between @berkeley_ai @Azure and @AIatMeta ! 🤝 Drawing parallels between LLMs and students in open-book (RAG) 📔 and closed-book exams (SFT) 🧠, we present a better recipe

6

49

231

1

7

Lisa Dunlap

@lisabdunlap

3 months

Stop by at arch exhibit hall 102 today to chat about our work in describing differences in image sets! And if you have some sets you want deceived we will be taking requests : D

Yuhui Zhang

@Zhang_Yu_hui

3 months

Excited to be at #CVPR2024 in Seattle! We will be presenting: 1. Describing Differences Between Two Image Sets (w/ @lisabdunlap ): Oral: Friday 1-2:30 PM @ Summit Hall C Poster: - Tuesday (tomorrow) 9:40-11 AM @ Arch Exhibit Hall #102 - Friday 5-6:30 PM @ Arch 4A-E #115 2.

1

4

23

0

8

Lisa Dunlap

@lisabdunlap

9 months

@sarahookr Still recovering from the workshop panel: “What is overhyped? Muchanistic interpretability” :p

0

7

Lisa Dunlap

@lisabdunlap

4 months

I <3 Llama 3

lmsys.org

@lmsysorg

4 months

Exciting new blog -- What’s up with Llama-3? Since Llama 3’s release, it has quickly jumped to top of the leaderboard. We dive into our data and answer below questions: - What are users asking? When do users prefer Llama 3? - How challenging are the prompts? - Are certain users

14

120

744

0

1

7

Lisa Dunlap

@lisabdunlap

5 months

If you like VisDiff, there’s some other great work on discovering biases in AI systems also at CVPR!

Sangwoo Mo

@sangwoomo

5 months

Biases in vision models pose a critical threat to the deployment of AI systems. How can we detect potential issues (semi-)automatically? #CVPR2024 highlight

1

7

29

0

6

Lisa Dunlap

@lisabdunlap

11 months

dont forget the new (+improved) sky logo

Joey Gonzalez

@profjoeyg

11 months

My students decorated the Sky Computing Lab @UCBerkeley with a 🍭 candy land theme 🍬. Was this your doing @lisabdunlap ?

1

2

15

0

6

Lisa Dunlap

@lisabdunlap

11 months

Saw cool work at ICCV by @BrinkmannJannik on the impact of data, model type, and training objective on biase. Made me rethink my strong stance that "biased model = biased dataset". Perhaps the model itself actually does matter a lot...

A Multidimensional Analysis of Social Biases in Vision Transformers

The embedding spaces of image models have been shown to encode a range of social biases such as racism and sexism. Here, we investigate specific factors that contribute to the emergence of these...

arxiv.org

1

3

5

Lisa Dunlap

@lisabdunlap

1 year

@conor_power23 Also don’t forget Georgios Pavlakos going to UT Austin for computer vision! He’s THE guy of you want to do 3d vision

0

5

Lisa Dunlap

@lisabdunlap

6 months

My friend Suzie wrote a blog about chickens.

0

5

Lisa Dunlap

@lisabdunlap

10 months

Worlds coldest take, but gmail please add an ACK feature, it feels too much sending an email just saying “sounds good”

0

5

Lisa Dunlap

@lisabdunlap

9 months

[4/5] Some things we found with VisDiff: * ResNet trained on ImageNet does worse when a person is present * StableDiffusionV2 produces more images in pictures frames than V1 * ImageNetV2 (likely) contains more social media images than ImageNet

2

5

Lisa Dunlap

@lisabdunlap

1 year

Sitting in the computer vision bay in lab often feels like I’m on the Big Bang Theory: one-liners followed by thunderous laughter every 45 seconds or so.

1

0

5

Lisa Dunlap

@lisabdunlap

10 months

@sh_reya I really struggle with this. To some degree I think working in a large field which values novelty needs to come with some pressure, but I should feel excited to read a paper about a topic I clearly am interested in. If anyone has this mindset I would love some wisdom.

2

1

5

Lisa Dunlap

@lisabdunlap

9 months

This is the most excited I have been about a new model in a while, amazing work Yutong!

Yutong Bai

@YutongBAI1002

9 months

How far can we go with vision alone? Excited to reveal our Large Vision Model! Trained with 420B tokens, effective scalability, and enabling new avenues in vision tasks! (1/N) Kudos to @younggeng @Karttikeya_m @_amirbar , @YuilleAlan Trevor Darrell @JitendraMalikCV Alyosha Efros!

18

160

1K

0

5

Lisa Dunlap

@lisabdunlap

11 months

I love the commitment to keeping the item “as is”.

1

0

5

Lisa Dunlap

@lisabdunlap

10 months

@IanArawjo Ah yes, I had this happen a few months ago…

0

5

Lisa Dunlap

@lisabdunlap

11 months

Tried using GPT-4V to generate latex code for my plots given a reference style. Got it wrong at first but I gave it a pic of the incorrect plot and it corrected its mistakes (although not perfect). Next I need to see if it can generate code to match my seaborn plots...

2

1

3

Lisa Dunlap

@lisabdunlap

5 months

@sarahookr @sarahookr @fffffelipec made this just for you :p

lmsys.org

@lmsysorg

5 months

🔥Exciting news -- GPT-4-Turbo has just reclaimed the No. 1 spot on the Arena leaderboard again! Woah! We collect over 8K user votes from diverse domains and observe its strong coding & reasoning capability over others. Hats off to @OpenAI for this incredible launch! To offer

52

202

1K

2

0

4

Lisa Dunlap

@lisabdunlap

6 months

0

4

Lisa Dunlap

@lisabdunlap

1 year

@PreetumNakkiran The ambiguity of the term out of distribution really irks me. Someone could show me a test example and my arguments for it being in or out of distribution would be equally valid

1

0

4

Lisa Dunlap

@lisabdunlap

9 months

[2/5] We explore a task we dub set difference captioning: given two image sets A and B, output language descriptions which are more true for A than B. To evaluate VisDiff we construct VisDiffBench, a benchmark of 187 paired image sets with ground-truth difference.

1

4

Lisa Dunlap

@lisabdunlap

6 months

It's concerning how much my high school self is similar to an LLM: when prompted incorrectly, we get right answer but then overthink change it at the last minute. Anyways some very cool work by very cool people!

Danny Halawi

@dannyhalawi15

6 months

Language models can imitate patterns in prompts. But this can lead them to reproduce inaccurate information if present in the context. Our work () shows that when given incorrect demonstrations for classification tasks, models first compute the correct

4

15

89

1

0

4

Lisa Dunlap

@lisabdunlap

9 months

[5/5] This work wouldn't be possible without the amazing collaboration between @Zhang_Yu_hui @wxh1996111 @syeung10 at @StanfordAILab and @ZhongRuiqi @trevordarrell @JacobSteinhardt @profjoey at @berkeley_ai . Paper: Code:

GitHub - Understanding-Visual-Datasets/VisDiff: Official implementation of "Describing Differences...

Official implementation of "Describing Differences in Image Sets with Natural Language" (CVPR 2024 Oral) - Understanding-Visual-Datasets/VisDiff

github.com

0

1

4

Lisa Dunlap

@lisabdunlap

11 months

@profjoeyg @UCBerkeley i will say i cannot take full responsibility for this masterpiece, @conor_power23 + many other students were also down to clown

0

3

Lisa Dunlap

@lisabdunlap

9 months

To @Wenliang_Dai and other InstructBLIP authors: why do so many captions get the first few sentences right but then hallucinate by saying something like "In addition, there are several other X scattered throughout the scene" Ex:

clipinvariance

Weights & Biases, developer tools for machine learning

wandb.ai

1

0

3

Lisa Dunlap

@lisabdunlap

1 year

@conor_power23 Jealous of their soon-to-be students, these people are truly fantastic!

0

2

Lisa Dunlap

@lisabdunlap

11 months

@brjathu well played my friend.

0

3

Lisa Dunlap

@lisabdunlap

9 months

[3/5] We create VisDiff, a super stellar system that proposes and score differences given two sets, returning a list of difference descriptions sorted by how well they differentiate the two groups.

1

3

Lisa Dunlap

@lisabdunlap

11 months

At the British museum and I really don’t understand why everyone here looks so calm; the first room I walked into had a wall of Greek pottery from the 4th century BCE presented like it was no big thing and I have been losing my mind since then

0

3

Lisa Dunlap

@lisabdunlap

4 months

Learn about how humans interact with LLMs and win $25k in the process... win-win if you ask me

lmsys.org

@lmsysorg

4 months

Exciting news -- we're thrilled to announce that LMSYS + @kaggle are launching a human preference prediction competition with $100,000 in prizes! Your challenge is to predict which responses users will prefer in head-to-head battles between LLMs in the Chatbot Arena real-world

8

69

491

0

3

Lisa Dunlap

@lisabdunlap

6 months

Also check out Tianjun’s podcast episode!

#7 - Tianjun Zhang: building LLM's that connect to the real world and...

Tianjun Zhang is a 6th year PhD student at UC Berkeley working on RL, LLMs, and other buzzwords related to ML hype. He also has six guinea pigs and laughs ev...

www.youtube.com

2

1

3

Lisa Dunlap

@lisabdunlap

4 months

Absolutely fascinating conversations, I didn’t think traffic law could be so philosophical

Conor Power

@conor_power23

4 months

New podcast ep with @lgrinberg talking about his transition from software engineer to lawyer! Very interesting to learn about how different the legal industry is from tech :)

0

5

0

3

Lisa Dunlap

@lisabdunlap

6 months

Congrats on another positively perfect project my friend :)

Shishir Patil

@shishirpatil_

6 months

📢Excited to release the live Berkeley Function-Calling Leaderboard! 🔥 Also debuting openfunctions-v2 🤩 the latest open-source SoTA function-calling model on-par with GPT-4🆕Native support for Javascript, Java, REST! 🫡 Leaderboard: Blog:

9

68

300

0

3

Lisa Dunlap

@lisabdunlap

11 months

@conor_power23 Ayy that was my first conference too! Very good vibe and fun people

0

2

Lisa Dunlap

@lisabdunlap

9 months

@conor_power23 Here’s my pitch: rent out a carnival cruise ship to hold the next ML conference

1

0

2

Lisa Dunlap

@lisabdunlap

11 months

When your safety training was too effective..

1

0

2

Lisa Dunlap

@lisabdunlap

11 months

@conor_power23 Do the Adam Sandler method: make all your fun lab mates co authors

1

0

2

Lisa Dunlap

@lisabdunlap

1 year

@eternalroree @conor_power23 I would argue it’s more of an epidemic than a bug

0

2

Lisa Dunlap

@lisabdunlap

11 months

@conor_power23 @justinjaffray Be the change that you want to see in the world, as they say

1

0

2

Lisa Dunlap

@lisabdunlap

11 months

Anyone have a good universal model for finding subtle differences in two images? I found quite a few papers on change detection but they all seem to be task specific. I tried feeding in two images to LLaVa side-by-side but it isn't looking too promising...

1

0

1

Lisa Dunlap

@lisabdunlap

11 months

@pecey01 For figures, start with a very simple version and have 5-10 people look at it who have only read your abstract. I have wasted many hours on complex figures that no one understands. Similarly, consider adding a teaser fig which just defines inputs/outputs.

0

2

Lisa Dunlap

@lisabdunlap

11 months

Good point from @profjoeyg , the feedback image may not have been used since I said the graph was cut off…

0

2

Lisa Dunlap

@lisabdunlap

11 months

Special thanks to @enfleisig and @YutongBAI1002 for curating the vibe

0

2

Lisa Dunlap

@lisabdunlap

1 year

@beenwrekt Via Rorschach rules, the number is actually dependent on the personality traits of the observer.

0

2

Lisa Dunlap

@lisabdunlap

1 year

Welcoming in the spooky season with @genmoai 's new video model

2

0

2

Lisa Dunlap

@lisabdunlap

1 year

this one is just cool

0

2

Lisa Dunlap

@lisabdunlap

9 months

I will be requiring all those who visit to sign my poster tube

0

2

Lisa Dunlap

@lisabdunlap

11 months

@profjoeyg @UCBerkeley Maybe……

1

0

1

Lisa Dunlap

@lisabdunlap

4 months

@KenAKAFrosty I haven't noticed this from my analysis (im-also-a-good-gpt2-chatbot performance on creative writing at least was quite good, same with creative writing for longer queries), but it would be interesting to look into this further

1

0

1

Lisa Dunlap

@lisabdunlap

1 year

WandB stop manipulating me.

0

1

Lisa Dunlap

@lisabdunlap

9 months

@Zhang_Yu_hui @berkeley_ai We even both went to @ZhongRuiqi separately and asked for feedback lol

0

1

Lisa Dunlap

@lisabdunlap

7 months

@deliprao @profjoeyg @beenwrekt warning: it's still in the cringe stage

1

0

Lisa Dunlap

@lisabdunlap

5 months

@sarahookr ...yes we do...

0

Lisa Dunlap

@lisabdunlap

11 months

@sh_reya are you sure this person wasn’t me? :p

0

1

Lisa Dunlap

@lisabdunlap

11 months

@pecey01 So many, but the most TL;DR version is read one of the many medium articles on how to make plots look good (i am also partial to the Paired color palette).

0

1

Lisa Dunlap

@lisabdunlap

6 months

@sh_reya @conor_power23 Top tier guest might I add

0

1

Lisa Dunlap

@lisabdunlap

11 months

@profjoeyg @charlespacker amazing ep but we really need a better camera setup in lab...

0

1

Lisa Dunlap

@lisabdunlap

9 months

@sarahookr Highly recommend the Lindsey Lohan Christmas movie on Netflix if you love Hallmark-style movies of questionable quality

0

Lisa Dunlap

@lisabdunlap

9 months

@untitled01ipynb @lisabdunlap - I want to flex on my labs meme slack channel

1

0

1

Lisa Dunlap

@lisabdunlap

9 months

@lateinteraction What can I say, it’s my favorite open source project

1

0

1

Lisa Dunlap

@lisabdunlap

1 year

another eerie one

0

1

Lisa Dunlap

@lisabdunlap

10 months

@sh_reya Agreed. I found a bug last night because I chose to visualize the images… I think a good 25% of my bugs are caught by looking at data/predictions

0

1

Lisa Dunlap

@lisabdunlap

9 months

@sangwoomo I’d love to chat!

0

1

Lisa Dunlap

@lisabdunlap

6 months

@shmu_h Apologies on my sound quality, we are still working on it...

0

1

Lisa Dunlap

@lisabdunlap

5 months

@profjoeyg And vibe based evals!

1

0

1

Lisa Dunlap

@lisabdunlap

10 months

@joe_hellerstein @sh_reya @lintool Thank you for supportive words. Luckily research at Berkeley is so dang fun that even with the anxiety inducing slack messages it’s well worth it :)

0

1

Lisa Dunlap

@lisabdunlap

5 months

Also shoutout all of our amazing co-authors! @wxh1996111 @ZhongRuiqi @trevordarrell @JacobSteinhardt @profjoeyg @syeung10

0

1

Lisa Dunlap

@lisabdunlap

7 months

I still don't know who was right, nor do I know if I understand how images work

0

1

Lisa Dunlap

@lisabdunlap

7 months

@deliprao @profjoeyg @beenwrekt We also have a podcast where grad students gossip:

Thinking About Computers • A podcast on Spotify for Podcasters

Two CS grad students talk to other researchers about grad school and anything else. Note: Please excuse any sniffling or snorting, Lisa has Tourette's and we haven't figured out a good workaround yet.

podcasters.spotify.com

1

0

1