\n\t\t\n\t\n\t\n\t\tNemotron-4 15B: Exploring the Power of a Cutting-Edge Multilingual Model\n\t\n\n

\n\n

\n\t\n\t\t\n\t\n\t\n\t\tLinks 🔗:\n\t\n

👉 Subscribe: https://www.youtube.com/@Arxflix
👉 Twitter: https://x.com/arxflix
👉 LMNT (Partner): https://lmnt.com/

\n","updatedAt":"2024-06-08T18:51:06.403Z","author":{"_id":"6186ddf6a7717cb375090c01","avatarUrl":"/avatars/716b6a7d1094c8036b2a8a7b9063e8aa.svg","fullname":"Julien BLANCHON","name":"blanchon","type":"user","isPro":true,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":176,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.45611852407455444},"editors":["blanchon"],"editorAvatarUrls":["/avatars/716b6a7d1094c8036b2a8a7b9063e8aa.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2402.16819","authors":[{"_id":"65dd6be65012ec503f1bae87","user":{"_id":"63d2f8182727d7888cba3c93","avatarUrl":"/avatars/fa46bab692fcf20b92270ec0fff3681a.svg","isPro":false,"fullname":"Jupinder Parmar","user":"jupinderp","type":"user"},"name":"Jupinder Parmar","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:38:03.076Z","hidden":false},{"_id":"65dd6be65012ec503f1bae88","user":{"_id":"6470efe7fa9fd77212e17a97","avatarUrl":"/avatars/a116b279b927627eeee9c23021b6e5ee.svg","isPro":false,"fullname":"Shrimai Prabhumoye","user":"shrimai","type":"user"},"name":"Shrimai Prabhumoye","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:38:09.023Z","hidden":false},{"_id":"65dd6be65012ec503f1bae89","user":{"_id":"645ad6e2d8ba048d02ad374a","avatarUrl":"/avatars/461a871387bbc0b3e117fd17dfce5430.svg","isPro":false,"fullname":"Joseph Jennings","user":"jojennin","type":"user"},"name":"Joseph Jennings","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:38:16.476Z","hidden":false},{"_id":"65dd6be65012ec503f1bae8a","user":{"_id":"630544b09d2531fabd156fd3","avatarUrl":"/avatars/7b5374244a887577834fb4524ff76d01.svg","isPro":false,"fullname":"Mostofa Patwary","user":"mpatwary","type":"user"},"name":"Mostofa Patwary","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:38:23.518Z","hidden":false},{"_id":"65dd6be65012ec503f1bae8b","user":{"_id":"627bf27cf19c5eb46d54cea8","avatarUrl":"/avatars/b8ca0b4e841858c1d234671187234f56.svg","isPro":false,"fullname":"Sandeep Subramanian","user":"MaximumEntropy","type":"user"},"name":"Sandeep Subramanian","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:38:30.421Z","hidden":false},{"_id":"65dd6be65012ec503f1bae8c","user":{"_id":"67e34615c15564fd97f697ce","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/BNXjsEA5rGn7uFoFuVkB6.png","isPro":false,"fullname":"Dan Su","user":"sudandandansu1","type":"user"},"name":"Dan Su","status":"claimed_verified","statusLastChangedAt":"2025-06-05T07:31:24.099Z","hidden":false},{"_id":"65dd6be65012ec503f1bae8d","name":"Chen Zhu","hidden":false},{"_id":"65dd6be65012ec503f1bae8e","user":{"_id":"617af1457ced0304d1b75b89","avatarUrl":"/avatars/5fdf3b5d9b93af1a70c4a81eb18a8d19.svg","isPro":false,"fullname":"Deepak Narayanan","user":"deepakn94","type":"user"},"name":"Deepak Narayanan","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:39:09.758Z","hidden":false},{"_id":"65dd6be65012ec503f1bae8f","user":{"_id":"63a1eeecf30c4642277e613e","avatarUrl":"/avatars/50d0fc6f0ce80bb8905b28e4fbe8cadb.svg","isPro":false,"fullname":"Aastha Jhunjhunwala","user":"aasthaj","type":"user"},"name":"Aastha Jhunjhunwala","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:39:15.893Z","hidden":false},{"_id":"65dd6be65012ec503f1bae90","user":{"_id":"65cc27efa994380c82ca603c","avatarUrl":"/avatars/651e2fbc272b1e8f8cd343c4c5ff6276.svg","isPro":false,"fullname":"Ayush Dattagupta","user":"ayushdg95","type":"user"},"name":"Ayush Dattagupta","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:39:22.510Z","hidden":false},{"_id":"65dd6be65012ec503f1bae91","user":{"_id":"646ac857086023e36edcc273","avatarUrl":"/avatars/3b6e51d9ff311bb472be98d5cdc947da.svg","isPro":false,"fullname":"Vibhu Jawa","user":"VibhuJawa","type":"user"},"name":"Vibhu Jawa","status":"claimed_verified","statusLastChangedAt":"2024-02-27T16:08:15.225Z","hidden":false},{"_id":"65dd6be65012ec503f1bae92","user":{"_id":"627fbbe9b846ecb5f28e6d0c","avatarUrl":"/avatars/733e198097cfa66c2300e471101735ff.svg","isPro":false,"fullname":"Jiwei Liu","user":"daxiongshu","type":"user"},"name":"Jiwei Liu","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:39:50.905Z","hidden":false},{"_id":"65dd6be65012ec503f1bae93","user":{"_id":"6362e35db64a5669426c4dfc","avatarUrl":"/avatars/430eafb1a5951480f7edf5735d72709e.svg","isPro":false,"fullname":"Ameya Mahabaleshwarkar","user":"amahabal","type":"user"},"name":"Ameya Mahabaleshwarkar","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:40:06.314Z","hidden":false},{"_id":"65dd6be65012ec503f1bae94","name":"Osvald Nitski","hidden":false},{"_id":"65dd6be65012ec503f1bae95","name":"Annika Brundyn","hidden":false},{"_id":"65dd6be65012ec503f1bae96","user":{"_id":"6556621866423b57b2ce0552","avatarUrl":"/avatars/de80b5d50ef4f84504ebc5f39c6bf147.svg","isPro":true,"fullname":"James Maki","user":"jmaki","type":"user"},"name":"James Maki","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:40:49.000Z","hidden":false},{"_id":"65dd6be65012ec503f1bae97","user":{"_id":"61900ccbec759925a87f919b","avatarUrl":"/avatars/1ba5b639fb871da839d6ad7db427755a.svg","isPro":false,"fullname":"Miguel","user":"miguelusque","type":"user"},"name":"Miguel Martinez","status":"claimed_verified","statusLastChangedAt":"2024-02-28T08:28:16.328Z","hidden":false},{"_id":"65dd6be65012ec503f1bae98","name":"Jiaxuan You","hidden":false},{"_id":"65dd6be65012ec503f1bae99","name":"John Kamalu","hidden":false},{"_id":"65dd6be65012ec503f1bae9a","name":"Patrick LeGresley","hidden":false},{"_id":"65dd6be65012ec503f1bae9b","user":{"_id":"64ff36132f27723840f0490d","avatarUrl":"/avatars/e2b559c81b1892a0e61aae5b9d08133c.svg","isPro":false,"fullname":"Denys Fridman","user":"dfridman","type":"user"},"name":"Denys Fridman","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:41:40.930Z","hidden":false},{"_id":"65dd6be65012ec503f1bae9c","name":"Jared Casper","hidden":false},{"_id":"65dd6be65012ec503f1bae9d","name":"Ashwath Aithal","hidden":false},{"_id":"65dd6be65012ec503f1bae9e","user":{"_id":"622937a4acd5bef90e55c49d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1649475932211-622937a4acd5bef90e55c49d.jpeg","isPro":false,"fullname":"Oleksii Kuchaiev","user":"okuchaiev","type":"user"},"name":"Oleksii Kuchaiev","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:41:55.083Z","hidden":false},{"_id":"65dd6be65012ec503f1bae9f","user":{"_id":"6641544c695975af2cbd0da6","avatarUrl":"/avatars/0ad3c18dcba585259b064fe9b00a07ce.svg","isPro":false,"fullname":"Mohammad Shoeybi","user":"shoeybi","type":"user"},"name":"Mohammad Shoeybi","status":"claimed_verified","statusLastChangedAt":"2024-05-28T07:19:41.371Z","hidden":false},{"_id":"65dd6be65012ec503f1baea0","name":"Jonathan Cohen","hidden":false},{"_id":"65dd6be65012ec503f1baea1","user":{"_id":"6311021788942700629e6247","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6311021788942700629e6247/vDktfPZwSRRhYIxkUKYjP.jpeg","isPro":false,"fullname":"Bryan Catanzaro","user":"ctnzr","type":"user"},"name":"Bryan Catanzaro","status":"admin_assigned","statusLastChangedAt":"2024-02-27T10:42:34.450Z","hidden":false}],"publishedAt":"2024-02-26T18:43:45.000Z","submittedOnDailyAt":"2024-02-27T02:28:15.440Z","title":"Nemotron-4 15B Technical Report","submittedOnDailyBy":{"_id":"60f1abe7544c2adfd699860c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674929746905-60f1abe7544c2adfd699860c.jpeg","isPro":false,"fullname":"AK","user":"akhaliq","type":"user"},"summary":"We introduce Nemotron-4 15B, a 15-billion-parameter large multilingual\nlanguage model trained on 8 trillion text tokens. Nemotron-4 15B demonstrates\nstrong performance when assessed on English, multilingual, and coding tasks: it\noutperforms all existing similarly-sized open models on 4 out of 7 downstream\nevaluation areas and achieves competitive performance to the leading open\nmodels in the remaining ones. Specifically, Nemotron-4 15B exhibits the best\nmultilingual capabilities of all similarly-sized models, even outperforming\nmodels over four times larger and those explicitly specialized for multilingual\ntasks.","upvotes":46,"discussionId":"65dd6be75012ec503f1baec1","ai_summary":"Nemotron-4 15B, a large multilingual language model, excels in English, multilingual, and coding tasks, demonstrating superior performance in multilingual capabilities compared to larger and specialized models.","ai_keywords":["large multilingual language model","downstream evaluation areas","multilingual capabilities"]},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"635cada2c017767a629db012","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1667018139063-noauth.jpeg","isPro":false,"fullname":"Ojasvi Singh Yadav","user":"ojasvisingh786","type":"user"},{"_id":"6538119803519fddb4a17e10","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6538119803519fddb4a17e10/ffJMkdx-rM7VvLTCM6ri_.jpeg","isPro":false,"fullname":"samusenps","user":"samusenps","type":"user"},{"_id":"62b20f6df4a72794189248fc","avatarUrl":"/avatars/87e1125868616d4f7d6ee1e5ec4499b4.svg","isPro":false,"fullname":"Ethan He","user":"ethanhe","type":"user"},{"_id":"64747f7e33192631bacd8831","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64747f7e33192631bacd8831/dstkZJ4sHJSeqLesV5cOC.jpeg","isPro":false,"fullname":"Taufiq Dwi Purnomo","user":"taufiqdp","type":"user"},{"_id":"6460a775051604bda029eaeb","avatarUrl":"/avatars/ec806d55e3a319542bbcfc5eea72b244.svg","isPro":false,"fullname":"Michael Ruppert","user":"miweru","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"61b85ce86eb1f2c5e6233736","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1655385361868-61b85ce86eb1f2c5e6233736.jpeg","isPro":false,"fullname":"Vaibhav Srivastav","user":"reach-vb","type":"user"},{"_id":"65b3c65066faac7b78e22e3e","avatarUrl":"/avatars/a59b1991366df8552835851dd707bcb0.svg","isPro":false,"fullname":"Aljoscha Grunwald","user":"roquueee","type":"user"},{"_id":"646ac857086023e36edcc273","avatarUrl":"/avatars/3b6e51d9ff311bb472be98d5cdc947da.svg","isPro":false,"fullname":"Vibhu Jawa","user":"VibhuJawa","type":"user"},{"_id":"6556621866423b57b2ce0552","avatarUrl":"/avatars/de80b5d50ef4f84504ebc5f39c6bf147.svg","isPro":true,"fullname":"James Maki","user":"jmaki","type":"user"},{"_id":"65cc27efa994380c82ca603c","avatarUrl":"/avatars/651e2fbc272b1e8f8cd343c4c5ff6276.svg","isPro":false,"fullname":"Ayush Dattagupta","user":"ayushdg95","type":"user"},{"_id":"6478281d4ae93470ffc6f2f8","avatarUrl":"/avatars/0b12b5e9e78cea782b12b5cec454909b.svg","isPro":false,"fullname":"Sanjeev Satheesh","user":"issanjeev","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":2}">

Papers

arxiv:2402.16819

Nemotron-4 15B Technical Report

Published on Feb 26, 2024

· Submitted by

AK on Feb 27, 2024

#2 Paper of the day

Upvote

Authors:

Jupinder Parmar ,

Shrimai Prabhumoye ,

Joseph Jennings ,

Mostofa Patwary ,

Sandeep Subramanian ,

Dan Su ,

Deepak Narayanan ,

Aastha Jhunjhunwala ,

Ayush Dattagupta ,

Vibhu Jawa ,

Jiwei Liu ,

Ameya Mahabaleshwarkar ,

James Maki ,

Miguel Martinez ,

Denys Fridman ,

Abstract

Nemotron-4 15B, a large multilingual language model, excels in English, multilingual, and coding tasks, demonstrating superior performance in multilingual capabilities compared to larger and specialized models.

AI-generated summary

We introduce Nemotron-4 15B, a 15-billion-parameter large multilingual language model trained on 8 trillion text tokens. Nemotron-4 15B demonstrates strong performance when assessed on English, multilingual, and coding tasks: it outperforms all existing similarly-sized open models on 4 out of 7 downstream evaluation areas and achieves competitive performance to the leading open models in the remaining ones. Specifically, Nemotron-4 15B exhibits the best multilingual capabilities of all similarly-sized models, even outperforming models over four times larger and those explicitly specialized for multilingual tasks.

View arXiv page View PDF Add to collection