Omri Abend (@abendomri) 's Twitter Profile
Omri Abend

@abendomri

Faculty member at the Hebrew University working on NLP/CL.

ID: 1087260366340583427

linkhttp://www.cs.huji.ac.il/~oabend/ calendar_today21-01-2019 08:07:26

409 Tweet

292 Followers

178 Following

Eitan Wagner (@eitanwagner) 's Twitter Profile Photo

Language Models output probabilities for tokens. So probabilities of spans must follow a valid joint, right? Check out our new #EMNLP2024 paper -- "CONTESTS: a Framework for Consistency Testing of Span Probabilities in Language Models". w/ Yuli Slavutsky and Omri Abend (1/5)

Eitan Wagner (@eitanwagner) 's Twitter Profile Photo

Language Models learn data with complex structures. Can they learn simple ones? According to our new #EMNLP2024 paper โ€œExploring the Learning Capabilities of Language Models using LEVERWORLDSโ€, the answer is โ€œYes, but not so fastโ€ฆโ€ w/ Amir Feder and Omri Abend (1/5)

๐ŸŽ— Naama Lazimi - ื ืขืžื” ืœื–ื™ืžื™ (@naamalazimi) 's Twitter Profile Photo

ื”ืฉื™ื— ืฉืœ ื™ืฉืจืืœ ื”ืจืืฉื•ื ื” ื•ื”ืฉื ื™ื” "ื—ื–ืจ ืืœื™ื ื•" ืื– ืจื’ืข ืœืคื ื™ ืฉื ืขื™ืฃ ืื•ืชื• ื•ืืช ืžื•ื‘ื™ืœื™ื• ืฉืขืฉื• ื‘ื ื• ืฉืžื•ืช, ื”ื™ื” ืœื™ ืžื” ืœื•ืžืจ >>

ื”ืกื•ืœื™ื“ื™ืช (@hasolidit) 's Twitter Profile Photo

ื“ื•ื‘ืจ ืฆื”ืดืœ ืืคื™ ื“ืคืจื™ืŸ - Effie Defrin ืืžืจืช ืืช ื”ืืžืช. ืืชื” ื’ื™ื‘ื•ืจ ื™ืฉืจืืœ. ืœื—ืžืช ื•ื’ื‘ืจืช ืขืœ ืื•ื™ื‘ื™ื ืงืฉื™ื ื™ื•ืชืจ ืžืœื”ืงืช ื”ืคื™ืจืื ื•ืช ื”ืžืชืงืจืืช ืžืžืฉืœืช ื”-7 ื‘ืื•ืงื˜ื•ื‘ืจ ืขืœ ืžืฉืชืžื˜ื™ื”, ื—ืฉื•ื“ื™ื”, ื ืืฉืžื™ื” ื•ืขื‘ืจื™ื™ื ื™ื”.

Esther Shizgal (@esthershizgal) 's Twitter Profile Photo

Happy to share that our paper is now on Arxiv ๐Ÿค— ๐Ÿšž Applying NLP for analyzing character development ๐Ÿ‘ฃ โœก๏ธ Examining Religious Trajectories in 1000 Holocaust testimonies ๐Ÿ•Ž ๐Ÿ‚๐ŸŒฑ๐Ÿ‚๐ŸŒฑ๐ŸŒฑ Sequence Clustering ๐Ÿ‚๐ŸŒฑ๐Ÿ‚๐ŸŒฟ๐ŸŒฑ arxiv.org/abs/2412.17063

ื”ืื•ื ื™ื‘ืจืกื™ื˜ื” ื”ืขื‘ืจื™ืช ื‘ื™ืจื•ืฉืœื™ื (@hebrewu_heb) 's Twitter Profile Photo

ื”ืื•ื ื™ื‘ืจืกื™ื˜ื” ื”ืขื‘ืจื™ืช, ื™ื—ื“ ืขื ืื•ื ื™ื‘ืจืกื™ื˜ืื•ืช ื”ืžื—ืงืจ ื‘ื™ืฉืจืืœ: ืงืžืคื™ื™ืŸ "ื“ื•ืจ ื”ื ื™ืฆื—ื•ืŸ" ื‘ื• ื ื˜ืขืŸ ื›ื™ ื ืขืฉื™ืช ื”ืขื“ืคื” ืžืชืงื ืช ืฉืœ ืขืจื‘ื™ื ื‘ืงื‘ืœื” ืœืœื™ืžื•ื“ื™ื, ืขืœ ื—ืฉื‘ื•ืŸ ื—ื™ื™ืœื™ ื”ืžื™ืœื•ืื™ื- ื”ื•ื ืงืžืคื™ื™ืŸ ืคื•ืœื™ื˜ื™ ื”ืžื‘ื•ืกืก ืขืœ ื ืชื•ื ื™ื ืฉืงืจื™ื™ื. ื‘ืžื›ืชื‘ ื”ืชืจืื” ืœืคื ื™ ืฆืขื“ื™ื ืžืฉืคื˜ื™ื, ื›ืชื‘ื• ืจืืฉื™ ื”ืื•ื ื™ื‘ืจืกื™ื˜ืื•ืช ื›ื™: "ื‘ืคืจืกื•ื ืžื•ืฆื’ื•ืช

ื”ืื•ื ื™ื‘ืจืกื™ื˜ื” ื”ืขื‘ืจื™ืช, ื™ื—ื“ ืขื ืื•ื ื™ื‘ืจืกื™ื˜ืื•ืช ื”ืžื—ืงืจ ื‘ื™ืฉืจืืœ: ืงืžืคื™ื™ืŸ "ื“ื•ืจ ื”ื ื™ืฆื—ื•ืŸ" ื‘ื• ื ื˜ืขืŸ ื›ื™ ื ืขืฉื™ืช ื”ืขื“ืคื” ืžืชืงื ืช ืฉืœ ืขืจื‘ื™ื ื‘ืงื‘ืœื” ืœืœื™ืžื•ื“ื™ื, ืขืœ ื—ืฉื‘ื•ืŸ ื—ื™ื™ืœื™ ื”ืžื™ืœื•ืื™ื- ื”ื•ื ืงืžืคื™ื™ืŸ ืคื•ืœื™ื˜ื™ ื”ืžื‘ื•ืกืก ืขืœ ื ืชื•ื ื™ื ืฉืงืจื™ื™ื. 
ื‘ืžื›ืชื‘ ื”ืชืจืื” ืœืคื ื™ ืฆืขื“ื™ื ืžืฉืคื˜ื™ื, ื›ืชื‘ื• ืจืืฉื™ ื”ืื•ื ื™ื‘ืจืกื™ื˜ืื•ืช ื›ื™: "ื‘ืคืจืกื•ื ืžื•ืฆื’ื•ืช
AI21 Labs (@ai21labs) 's Twitter Profile Photo

Weโ€™re excited to announce that our new course, โ€˜Build Long Context AI Apps with Jambaโ€™, built in partnership with Andrew Ng DeepLearning.AI is now live - and itโ€™s currently free! In this course, Chen Wang and Chen Almagor from AI21 Labs will walk you through how to use

Uri Berger (@uriberger88) 's Twitter Profile Photo

1/ Caregivers teach infants how to speak by correcting errors. Child: "I goed to the park" Caregiver: "I went to the park" Can models do this too? Read our recent paper on reformulation in image captioning: arxiv.org/abs/2501.04513 w\ Omri Abend Lea Frermann Gabriel Stanovsky

1/ Caregivers teach infants how to speak by correcting errors.
Child: "I goed to the park"
Caregiver: "I went to the park"
Can models do this too? Read our recent paper on reformulation in image captioning:
arxiv.org/abs/2501.04513
w\
<a href="/AbendOmri/">Omri Abend</a>
<a href="/leafrermann/">Lea Frermann</a>
<a href="/GabiStanovsky/">Gabriel Stanovsky</a>
AI21 Labs (@ai21labs) 's Twitter Profile Photo

Excited to share that our paper, Jamba: Hybrid Transformer-Mamba Language Models, has been accepted to ICLR 2025! Weโ€™re honored to contribute to this important platform and the collective knowledge of the AI community by sharing our work openly and submitting it for peer review.

AI21 Labs (@ai21labs) 's Twitter Profile Photo

Today we launched Jamba 1.6, the best open model for private enterprise deployment. AI21โ€™s Jamba outperforms Cohere, Mistral and Llama on key benchmarks, including Arena Hard, and rivals leading closed models while maintaining unmatched speed and quality.ย  Now available on

Today we launched Jamba 1.6, the best open model for private enterprise deployment. AI21โ€™s Jamba outperforms Cohere, Mistral and Llama on key benchmarks, including Arena Hard, and rivals leading closed models while maintaining unmatched speed and quality.ย 

Now available on
Rafi reshef (@reshefrafi) 's Twitter Profile Photo

ื ืชื ื™ื”ื• ื”ื’ืจื™: ื—ืœื•ืงืช ืชืคืงื™ื“ื™ื ื‘ืจื•ืจื”

AI21 Labs (@ai21labs) 's Twitter Profile Photo

Meet Maestro: AI21 Labs new AI Planning and Orchestration system. Unlike LLM-based agents, #Maestro ensures reliable outcomes while optimizing latency and cost. This is AI that you can actually trust. ๐Ÿ”— Sign up now! ai21.com/maestro #AI21Maestro #AccuracyAtScale

CoNLL 2025 (@conll_conf) 's Twitter Profile Photo

๐Ÿšจ The CoNLL deadline is just 2 days away! ๐Ÿšจ Submit your work by March 14th, 11:59 PM (AoE, UTC-12) Don't miss out! โณ ๐Ÿ”— Submission links: conll.org #CoNLL2025 #NLP #CoNLL

Eitan Wagner (@eitanwagner) 's Twitter Profile Photo

- โ€œI flipped a biased coin with p(Heads) = 0.55.โ€ - โ€œWhat did it land on?โ€ What is the probability of the answer being โ€œHeadsโ€? Does it depend on whether the outcome is seen? Should we expect it to be 0.55? Check out our new paper! arxiv.org/abs/2505.02072 w/ Omri Abend (1/10)

Tomer Persico (@tomerpersico) 's Twitter Profile Photo

ืื ืืชื ืจื•ืฆื™ื ืœื“ืขืช ืœืžื” ื”ื™ืžื™ืŸ ืœื ืžืฆืœื™ื— ืœืžืฉื•ืœ ื’ื ืื—ืจื™ 50 ืฉื ื•ืช ืฉืœื˜ื•ืŸ ื›ืžืขื˜ ืจืฆื•ืฃ, ืงืจืื• ืืช ื”ื˜ื•ืจ ื”ืื—ืจื•ืŸ ืฉืœ ืงืœืžืŸ ืœื™ื‘ืกืงื™ื ื“ . ื”ื‘ืขื™ื” ื”ืขื™ืงืจื™ืช ืฉื”ื•ื ืžืžื—ื™ืฉ - ืœื ื‘ื˜ืงืกื˜ ืืœื ื‘ืกืื‘ื˜ืงืกื˜ - ื”ื™ื ื—ื•ืกืจ ื”ื™ื›ื•ืœืช ืœืคืชื— ืกื•ื›ื ื•ืช, ื•ื”ื™ืกื•ื“ ื”ืžืฉืœื™ื ืฉืœ ื–ื”, ืœืงื—ืช ืื—ืจื™ื•ืช. ืืคืฉืจ ืœืชืžืฆืช ืืช ื”ืžืืžืจ ื‘ืฆื™ื˜ื•ื˜ ื”ื‘ื: "ื›ืฉืื ื™ ืจื•ืื” ืืช

ืื ืืชื ืจื•ืฆื™ื ืœื“ืขืช ืœืžื” ื”ื™ืžื™ืŸ ืœื ืžืฆืœื™ื— ืœืžืฉื•ืœ ื’ื ืื—ืจื™ 50 ืฉื ื•ืช ืฉืœื˜ื•ืŸ ื›ืžืขื˜ ืจืฆื•ืฃ, ืงืจืื• ืืช ื”ื˜ื•ืจ ื”ืื—ืจื•ืŸ ืฉืœ ืงืœืžืŸ ืœื™ื‘ืกืงื™ื ื“ . ื”ื‘ืขื™ื” ื”ืขื™ืงืจื™ืช ืฉื”ื•ื ืžืžื—ื™ืฉ - ืœื ื‘ื˜ืงืกื˜ ืืœื ื‘ืกืื‘ื˜ืงืกื˜ - ื”ื™ื ื—ื•ืกืจ ื”ื™ื›ื•ืœืช ืœืคืชื— ืกื•ื›ื ื•ืช, ื•ื”ื™ืกื•ื“ ื”ืžืฉืœื™ื ืฉืœ ื–ื”, ืœืงื—ืช ืื—ืจื™ื•ืช.

ืืคืฉืจ ืœืชืžืฆืช ืืช ื”ืžืืžืจ ื‘ืฆื™ื˜ื•ื˜ ื”ื‘ื: 

"ื›ืฉืื ื™ ืจื•ืื” ืืช