Skip to page content
SAVRN
Search Plan a project

Dataset

dpo-icelandic-interpreted

by Jonas Lindh jonasaise/dpo-icelandic-interpreted

This dataset contains Direct Preference Optimization (DPO) pairs translated and culturally adapted into Icelandic from the original argilla/ultrafeedback-binarized-preferences-cleaned dataset.

Rows6,433
Configurations1
Size18.0 MB
Licensemit
AccessPublicly accessible
Monthly Downloads88

Dataset Card

By Jonas Lindh, published under mit, revision 2385c77cdc99.

Dataset Card for DPO Icelandic Interpreted

Dataset Description

This dataset contains Direct Preference Optimization (DPO) pairs translated and culturally adapted into Icelandic from the original argilla/ultrafeedback-binarized-preferences-cleaned dataset.

Origin and Methodology

  • Source Dataset: argilla/ultrafeedback-binarized-preferences-cleaned
  • Model Used for Synthesis: kimi-k3 (via local endpoint)
  • Translation Strategy: The dataset was generated using a specialized prompt designed to prevent "translationese" and direct literal translation. The kimi-k3 model was instructed to heavily localize the prompt and the chosen response to fit Icelandic culture and idioms. Furthermore, it intentionally generated a rejected response that suffers from poor translation, excessive English loan words (Swenglish/Ice-lish), or US-centric hallucinations. This provides a high-quality, targeted signal for DPO/RLHF to teach models to favor authentic, culturally grounded Icelandic over robotic dictionary translations.

Structure

The dataset contains the following fields: - prompt: The culturally adapted Icelandic prompt. - chosen: The high-quality, authentic Icelandic response. - rejected: The poorly translated, uncalibrated response. - language: The target language of the row.

License

This dataset inherits the license from its source dataset argilla/ultrafeedback-binarized-preferences-cleaned (MIT).

Structure

default 6,433 rows

SplitRowsSize
train6,43328.9 MB
promptlarge_stringchosenlarge_stringrejectedlarge_stringlanguagelarge_string

Details

Repository
jonasaise/dpo-icelandic-interpreted
Publisher
Jonas Lindh
Task category
Not stated by the source
Tags
dpo, rlhf, ultrafeedback
Size category
Not stated by the source
Languages
is
Revision
2385c77cdc993a4963244ebe116b4e387e362908
Last updated
2026-10-01

Files

131 files, 18.0 MB in total.

Data129 files · 18.0 MB
Documentation1 file · 1.7 KB
Repository1 file · 2.5 KB
Every file
FileTypeSizeSHA-256
data/train-1790100884.parquetData141.3 KB1289faf46f9d
data/train-1790101048.parquetData140.3 KBf712a4461142
data/train-1790102488.parquetData138.0 KB0d24397400fa
data/train-1790102719.parquetData159.1 KB46ea283007d5
data/train-1790104000.parquetData143.9 KB64f8150538a4
data/train-1790104125.parquetData120.3 KB97b4e4c5731c
data/train-1790104542.parquetData78.4 KB9a0f8024d312
data/train-1790104640.parquetData74.2 KBd3bbc000e01a
data/train-1790105050.parquetData68.9 KBdf86c0dcda6e
data/train-1790105157.parquetData75.5 KB4c4f2dba473e
data/train-1790105453.parquetData69.0 KB65b08d62702c
data/train-1790105659.parquetData72.8 KBa16257c75b14
data/train-1790106151.parquetData45.0 KB62299684acce
data/train-1790106180.parquetData26.0 KBf14ea1c0ab61
data/train-1790753998.parquetData138.9 KBc3fa61f5a328
data/train-1790754654.parquetData155.7 KB765c7cb0f255
data/train-1790755295.parquetData127.8 KBbda08d82ab1c
data/train-1790755886.parquetData128.9 KBd982c5136bee
data/train-1790756432.parquetData133.5 KB5eca64162274
data/train-1790757033.parquetData145.8 KBe5b50b7fbe18
data/train-1790757647.parquetData137.6 KB41098b3ab28f
data/train-1790758386.parquetData159.1 KBc841df5cb12e
data/train-1790759029.parquetData127.3 KBa5b0c2be55c0
data/train-1790759633.parquetData127.6 KB23a88fdf48c9
data/train-1790760354.parquetData149.2 KB6b6901dfc053
data/train-1790760872.parquetData136.6 KBd2e3c66997f8
data/train-1790761576.parquetData165.3 KBf7fb6b274d65
data/train-1790762248.parquetData115.4 KBf20b7e164126
data/train-1790762900.parquetData130.5 KB39370007f307
data/train-1790763507.parquetData149.9 KB1a1c6c8fcfba
data/train-1790764095.parquetData137.8 KB5f48ae93e96e
data/train-1790764776.parquetData147.4 KBfae9dc92a649
data/train-1790765308.parquetData134.9 KB979221eccb30
data/train-1790765912.parquetData134.3 KB05660a6d6410
data/train-1790766502.parquetData137.6 KBdafdc8317e84
data/train-1790767062.parquetData121.7 KB1375b68ac6fd
data/train-1790767704.parquetData138.3 KB6f529818e5c4
data/train-1790768504.parquetData160.2 KB590cf20b5f9b
data/train-1790769348.parquetData163.7 KBf9c8bdd96387
data/train-1790769905.parquetData138.5 KBc8ea4702e317
data/train-1790770511.parquetData151.8 KBa306be34100d
data/train-1790771174.parquetData142.7 KBc1d6d5a87f7b
data/train-1790771766.parquetData147.0 KB3edd1974b9ea
data/train-1790772493.parquetData157.2 KB549d53e8d81e
data/train-1790773016.parquetData141.0 KBaaf00f4e9bba
data/train-1790773695.parquetData138.6 KB092ace176fdb
data/train-1790774320.parquetData134.9 KB1043f6585adb
data/train-1790774947.parquetData154.9 KBb2466fb80be1
data/train-1790775568.parquetData145.4 KBad0bd2e71525
data/train-1790776298.parquetData152.1 KB92868999960a
data/train-1790777048.parquetData144.3 KB8aeda9ffa67d
data/train-1790777717.parquetData140.0 KB2daa62f3f88f
data/train-1790778300.parquetData161.7 KB2b17e220282c
data/train-1790778851.parquetData138.1 KB25d4ab8ae3ea
data/train-1790779488.parquetData129.2 KB71479de83e21
data/train-1790780100.parquetData154.0 KB2a5e73e48457
data/train-1790780791.parquetData154.7 KB36c0f0dffe62
data/train-1790781459.parquetData160.9 KB8f895a661c21
data/train-1790782125.parquetData146.6 KBca642c1d7cac
data/train-1790782747.parquetData141.8 KBc35a4ca8dd16
data/train-1790783396.parquetData156.4 KB3c5b6b452d60
data/train-1790784054.parquetData152.2 KB44fa5f7667e8
data/train-1790784713.parquetData150.8 KBe9fbc5719b9f
data/train-1790785322.parquetData147.0 KB0e256efe3d57
data/train-1790785961.parquetData135.9 KB2cdf85e5865b
data/train-1790786598.parquetData158.4 KB9f0893043f86
data/train-1790787258.parquetData162.7 KBd77da3fd1d22
data/train-1790787804.parquetData125.1 KB07cacf11a6de
data/train-1790788489.parquetData157.8 KBb54b772d2964
data/train-1790789140.parquetData139.6 KB8bdfa9de04bf
data/train-1790789782.parquetData155.0 KBa38a1f4d2fb3
data/train-1790790455.parquetData146.0 KB846f5abd9b0d
data/train-1790791063.parquetData155.5 KBece5ca4c8189
data/train-1790791671.parquetData133.2 KBd27f3a1f9a40
data/train-1790792364.parquetData157.4 KB5f0fe2c00987
data/train-1790792905.parquetData147.5 KBa9f055400aa9
data/train-1790793486.parquetData117.3 KB06091587b2fd
data/train-1790794102.parquetData143.7 KBa9e0c88ba07f
data/train-1790794768.parquetData162.6 KB5a074ed4dc46
data/train-1790795483.parquetData168.7 KB717f7ae60afe
data/train-1790796222.parquetData161.4 KB3d0f0d53d220
data/train-1790796856.parquetData147.0 KB0285f7693859
data/train-1790797424.parquetData140.2 KB9a9a92b5b76d
data/train-1790798061.parquetData123.6 KB6359fd975e64
data/train-1790798795.parquetData156.5 KB94beb0536ce6
data/train-1790799432.parquetData140.3 KB73d04bd872e8
data/train-1790800063.parquetData133.8 KBb2df280f97ee
data/train-1790800705.parquetData157.0 KB2ec6d3281cd0
data/train-1790801382.parquetData153.0 KBc4c1b4c2207c
data/train-1790802007.parquetData124.7 KB49ca6839d54e
data/train-1790802583.parquetData147.1 KB079e075a386e
data/train-1790803180.parquetData155.8 KB247f592ddaf7
data/train-1790803850.parquetData144.5 KB593f07d777f8
data/train-1790804480.parquetData144.4 KBddf861bc73ac
data/train-1790805042.parquetData135.6 KBb2608fcb7c7d
data/train-1790805613.parquetData137.3 KB268e7007d250
data/train-1790806330.parquetData158.7 KB9c22f5b2313b
data/train-1790806919.parquetData140.0 KBed327170e5b0
data/train-1790807460.parquetData146.1 KB251c412483e6
data/train-1790808148.parquetData151.3 KB5a217fed6133
data/train-1790808689.parquetData147.7 KBaf6539344ad8
data/train-1790809303.parquetData145.1 KBa39d7228de5e
data/train-1790810010.parquetData148.5 KBc7265d4a8cf0
data/train-1790810533.parquetData132.7 KB67ede34f140e
data/train-1790811186.parquetData165.3 KBc1057641b81d
data/train-1790811777.parquetData155.8 KB6cbfb1ffd561
data/train-1790812335.parquetData145.8 KB05ccec71ec98
data/train-1790812943.parquetData152.4 KBcc34740724af
data/train-1790813533.parquetData135.9 KB58d504125e23
data/train-1790814109.parquetData150.8 KB4d5a54d75765
data/train-1790814734.parquetData147.0 KBae07e50232f8
data/train-1790815336.parquetData133.7 KB28cbfe70132a
data/train-1790815957.parquetData136.3 KB7e15898ad5c0
data/train-1790816641.parquetData169.5 KBf816e7ace366
data/train-1790817281.parquetData140.8 KBa87654e560cc
data/train-1790817967.parquetData159.7 KB264142ff3c14
data/train-1790818550.parquetData158.3 KB394c15fe1959
data/train-1790819190.parquetData134.6 KB44ecaf93fc2a
data/train-1790819797.parquetData130.4 KB14e56c4b625f
data/train-1790820435.parquetData138.1 KBe9e9961ea16c
data/train-1790821103.parquetData149.0 KB8eddfea169f4
data/train-1790821739.parquetData147.7 KBcd52924290df
data/train-1790822423.parquetData148.7 KB31b81ba7b28d
data/train-1790823022.parquetData135.1 KBb3cd8433877e
data/train-1790823659.parquetData159.5 KB61f58113bc32
data/train-1790824187.parquetData135.0 KB3a7573946daa
data/train-1790824748.parquetData130.9 KB015ad68ba294
data/train-1790825365.parquetData146.6 KBcbb187fcaabf
data/train-1790826031.parquetData149.0 KBb16b689bd7da
README.mdDocumentation1.7 KB—
.gitattributesRepository2.5 KB—

License and Download

License
mit
Access
No access gate
Download from Jonas Lindh

Released by Jonas Lindh through its official repository on Hugging Face. Read the license.