Wenting Zhao @wzhao_nlp Twitter profile

Last Seen Profiles

@ShadiBartsch

@TrustlessState

@dominan_slave

@tandfonline

@AteoBreaking

@FremenWormrider

@Palmette_r

@1o0550o1

@3EiE4

@Z10esports

@bpthaber

@iagharanya

@themellowing

@TanyaTate

@GbbStorm

@barbiebqluv

@Ryan3Zhu

@RevRevRick

@BatsinChurches

@77nDm63Msskh7V

@madbacker56

@OCH1682890

@moniguzman

@moonfish7777

@RJP4UP_

@red_jeunes

@RealMacReport

@lame_Fido

@REALMadMax1960

@canal13

@REIDA_fumo1ban

@Benzaie_tgwtg

@BenjaminPDixon

@thequirkyempor1

@AgentM_Crypto

@arcanjo2873

Wenting Zhao

@wzhao_nlp

9 months

Hi friends, for a limited time, our WildChat project will provide free access to GPT-4 turbo: . For any data we collect during this period, we will give it back to the community 😊 Paper: Thank @ai2_mosaic for the generous support.

Chat-with-GPT4Turbo - a Hugging Face Space by yuntian-deng

huggingface.co

5

22

117

Wenting Zhao

@wzhao_nlp

7 months

I'm happy to share that WildChat has been accepted as a spotlight paper at #ICLR2024 ! Since the release of the dataset, it has been able to support so much research such as multi-turn conversation evaluation, cultural analysis, etc. We will release more data soon, stay tuned 💙

Yuntian Deng

@yuntiandeng

9 months

WildChat dataset is out @ai2_mosaic 🚀 Explore 650K user-ChatGPT interactions in the wild: 🔗 A huge shoutout to the team @wzhao_nlp @xiangrenNLP @jmhessel @clairecardie @YejinChoinka Fun fact: The ChatGPT/GPT-4 chatbot often thought it was GPT-3🤣

3

46

176

2

13

107

Wenting Zhao

@wzhao_nlp

7 months

shout-out to my amazing mentors @alsuhr @xiang_lorraine I would do another PhD if they were my advisors 🫶

Ai2

@allen_ai

7 months

We're delighted to extend hearty congratulations to the AI2 Outstanding Intern of the Year recipients for 2023! Henry Herzog, Wenting Zhao ( @wzhao_nlp ), and Yue Yang ( @YueYangAI ) made exceptional contributions to the team. Learn more about our winners:

3

10

94

5

2

52

Wenting Zhao

@wzhao_nlp

1 year

Heading to #ACL2023 🚀 My collaborators @megamor2 @billyuchenlin @michiyasunaga @aman_madaan @taoyds and I will be presenting a cutting-edge tutorial on Complex Reasoning in Natural Language - diving into recent methods for accurate, robust & trustworthy reasoning systems🤖 1/2

2

11

49

Wenting Zhao

@wzhao_nlp

8 months

I am presenting a few cool projects at #EMNLP2023 !!! 1. "Symbolic Planning and Code Generation for Grounded Dialogue" (SPC) led by @justintchiu . He has the best vision for human-robot optimal communication, and he's currently on the job market! He's my no.1 goto collaborator 😍

1

4

31

Wenting Zhao

@wzhao_nlp

9 months

The NLP community has created many instruction fine-tuning datasets by imagining what are the use cases of chatbots, but what do the users really ask them to do? Check out WildChat to find out 😍

Yuntian Deng

@yuntiandeng

9 months

WildChat dataset is out @ai2_mosaic 🚀 Explore 650K user-ChatGPT interactions in the wild: 🔗 A huge shoutout to the team @wzhao_nlp @xiangrenNLP @jmhessel @clairecardie @YejinChoinka Fun fact: The ChatGPT/GPT-4 chatbot often thought it was GPT-3🤣

3

46

176

0

2

29

Wenting Zhao

@wzhao_nlp

1 year

I will be also presenting our paper (w/ @justintchiu @clairecardie @srush_nlp ) on "Abductive Commonsense Reasoning Exploiting Mutually Exclusive Explanations", diving into unsupervised abductive reasoning by leveraging the mutual exclusiveness of explanations. Eager to chat💡

0

2

24

Wenting Zhao

@wzhao_nlp

8 months

Awww, thanks Jack! ♥️ I’m at EMNLP, happy to chat about reasoning, evaluation, imitation learning, and anything about graduate school 😍

jack morris

@jxmnop

8 months

here are two awesome researchers you should follow: @WeijiaShi2 at UW and @wzhao_nlp at Cornell!! some of their recent work: weijia shi ( @WeijiaShi2 ): - built INSTRUCTOR, the embedding model that lots of startups / companies use () - proposed a more

1

14

111

0

18

Wenting Zhao

@wzhao_nlp

8 months

Come meet and chat with your fellow reasoning researchers ☺️

Yiyuan Li @Summer

@YiyuanLi1

8 months

Overwhelmed massive works about reasoning but still feel lack of nutritions? @wzhao_nlp and I are hosting a BoF session regarding "Recipes in Building Language Reasoners" in #EMNLP2023 at Singapore @emnlpmeeting . 1/3

2

0

7

0

1

10

Wenting Zhao

@wzhao_nlp

10 months

@ABosselut I don't know about others, but our second workshop on Natural Language Reasoning and Structured Explanations will be held at ACL'24 (shameless self-promotion🤣)

0

7

Wenting Zhao

@wzhao_nlp

8 months

Thanks to all my awesome collaborators: @justintchiu @saujasv @dan_fried @derek @derekchen14 and my most supportive, helpful, visionary advisors: @clairecardie @srush_nlp

1

0

5

Wenting Zhao

@wzhao_nlp

8 months

@shauryr This is standard student F1 visa :)

0

5

Wenting Zhao

@wzhao_nlp

9 months

love LLMs as productivity tools

Andrej Karpathy

@karpathy

9 months

ChatGPT "Advanced Data Analysis" (which doesn't really have anything to do with data specifically) is an awesome tool for creating diagrams. I could probably code these diagrams myself, but it's soo much better to just sit back, and iterate in English. In this example, I was

123

769

5K

0

5

Wenting Zhao

@wzhao_nlp

8 months

2. "HOP, UNION, GENERATE (HUG): Explainable Multi-hop Reasoning without Rationale Supervision" RAG is used for retrieval-augmented generation, but RAG retrieves documents independently. By explicitly modeling the multi-hop relationship between documents, HUG beats RAG!

1

0

5

Wenting Zhao

@wzhao_nlp

8 months

3. "MiniChain: A Small Library for Large LMs" @srush_nlp made a really lightweight and easy-to-use library that allows people to quickly prototype software with LLMs in the loop. @justintchiu uses minichain to build his SPC dialogue system 🚀

1

0

4

Wenting Zhao

@wzhao_nlp

8 months

Please also check out @jxmnop 's paper on "Text Embeddings Reveal (Almost) As Much as Text" His paper totally changes my belief about embeddings, from which you can exactly reconstruct the text. The experiments in this paper blows my mind 🥳

0

4

Wenting Zhao

@wzhao_nlp

9 months

@srush_nlp Have you seen this one? "transformer decoders with a logarithmic number of decoding steps (w.r.t. the input length) push the limits of standard transformers only slightly, while a linear number of decoding steps adds a clear new ability..."

1

0

4

Wenting Zhao

@wzhao_nlp

10 months

@NirantK @arankomatsuzaki Thanks for your interest. We will release the dataset and model once we get internal clearance. Stay tuned 😊

0

3

Wenting Zhao

@wzhao_nlp

10 months

@yanaiela Can you also include the projects that were killed and didn't even make to submission stages? 🤔

1

0

3

Wenting Zhao

@wzhao_nlp

9 months

@jxmnop @Teknium1 @geepytee i don't know for sure obviously 😅 but discriminative models seem to be doing better at identifying bad requests from our recent work

0

3

Wenting Zhao

@wzhao_nlp

7 months

@niloofar_mire Please bug me a lot if you have questions ❤️

0

3

Wenting Zhao

@wzhao_nlp

10 months

@jxmnop lol i know how you feel, i can probably earn a million $ by managing people's wandb accounts to reduce their stress

0

3

Wenting Zhao

@wzhao_nlp

6 months

@ReviewAcl @koustavagoswami @soldni We received a meta-review that contains factual errors. It also only summarizes the pre-rebuttal reviews and misses corrections and acknowledgments from the reviewers made during the rebuttal. The way it lists bullet points also seems to be ChatGPT generated.

1

0

3

Wenting Zhao

@wzhao_nlp

7 months

@random_walker @maria_antoniak @aryaman2020 @lmsysorg We actually have evidence that users regularly use this external APIs given that otherwise the users would have to pay money out of their own pockets to use GPT-4. For example, there was an editor used our service to draft emails for a few times.

0

2

Wenting Zhao

@wzhao_nlp

8 months

@jmhessel oh nooo!! hope you feel better soon 🩹

0

2

Wenting Zhao

@wzhao_nlp

9 months

@andersonbcdefg @s_tworkowski @ai2_mosaic the link will be up next monday

0

2

Wenting Zhao

@wzhao_nlp

6 months

@yanaiela congrats Yanai!!! we'll celebrate with food when we run into each other next time

0

1

Wenting Zhao

@wzhao_nlp

6 months

@aman_madaan @LChoshen @rtk254 sorry i'm late to the party 🥹 just recovered from the chaotic ARR Feb. thanks for thinking of me

0

2

Wenting Zhao

@wzhao_nlp

9 months

@HJCH0 Congrats!!!

2

0

1

Wenting Zhao

@wzhao_nlp

9 months

@EdwardSun0909 We have more data + it is legal to use our data (ShareGPT collected user data without their consent😅)

1

0

1

Wenting Zhao

@wzhao_nlp

10 months

@BlackboxNLP What about a main conference paper with a focus on interpretability 😆

1

0

1

Wenting Zhao

@wzhao_nlp

8 months

@nitindominic5 It's just renewing and I've done my undergrad in the US 😂

0

1

Wenting Zhao

@wzhao_nlp

9 months

@EdwardSun0909 I am not sure about the legal consequences of training models on ShareGPT. It is possible that users pursue legal actions against ShareGPT because no consent was given (but since it is uploaded by an anonymous user, I'm not sure how this step can be carried out).

1

0

1

Wenting Zhao

@wzhao_nlp

8 months

@srush_nlp Your video still looks good in 2x speed😆

1

0

1

Wenting Zhao

@wzhao_nlp

10 months

@denny_zhou I think I make careless mistakes and I can correct those. Claiming people cannot self-correct more than LLMs requires rigorous study and set up rigorous premises.

0

1

Wenting Zhao

@wzhao_nlp

8 months

@liweijianglw fingers crossed!!!!!

1

0

1

Wenting Zhao

@wzhao_nlp

6 months

@ReviewAcl @koustavagoswami @soldni We reported the case to the ARR editors but have not heard back. Can you advise what we should do? Thank you!!!

0

1

Wenting Zhao

@wzhao_nlp

2 years

@jxmnop But I wish you a happy birthday!

1

0

1

Wenting Zhao

@wzhao_nlp

9 months

@EdwardSun0909 This paper recently addresses this issue:

SILO Language Models: Isolating Legal Risk In a Nonparametric Datastore

The legality of training language models (LMs) on copyrighted or otherwise restricted data is under intense debate. However, as we show, model performance significantly degrades if trained only on...

arxiv.org

1

0

1