Datasets:
corpus
stringclasses 3
values | category
stringclasses 3
values | dataset
stringclasses 9
values | task
stringclasses 27
values | prompt
stringclasses 22
values | model
stringclasses 48
values | ckpt_num
int64 1k
47.7k
| score
float64 0
85.7
|
|---|---|---|---|---|---|---|---|
Top
|
Reading comprehension
|
fquad
|
fquad_p1
|
{{context}}
Question : {{question}}
En prenant en compte le texte précédent, la réponse à la question est
|
2B
| 1,000
| 1.063
|
Top
|
Reading comprehension
|
fquad
|
fquad_p2
|
{{context}}
En répondant en référençant le texte précédent, {{question}}
|
2B
| 1,000
| 0.01
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p0
|
Texte : {{flores_passage}}
Question : {{question}}
A : {{mc_answer1}}
B : {{mc_answer2}}
C : {{mc_answer3}}
D : {{mc_answer4}}
Réponse :
|
2B
| 1,000
| 24.667
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p1
|
{{flores_passage}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Réponse :
|
2B
| 1,000
| 23.778
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Texte : {{flores_passage}}
Réponse :
|
2B
| 1,000
| 24.333
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p0
|
La phrase suivante est correcte grammaticalement:
|
2B
| 1,000
| 57.983
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
2B
| 1,000
| 57.983
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
2B
| 1,000
| 25.21
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p0
|
Context: {{context}}
|
2B
| 1,000
| 36.62
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p1
|
{{context}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Réponse :
|
2B
| 1,000
| 18.31
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Texte : {{context}}
Réponse :
|
2B
| 1,000
| 14.085
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p0
|
La phrase suivante est logique sémantiquement:
|
2B
| 1,000
| 45.378
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
2B
| 1,000
| 48.739
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
2B
| 1,000
| 15.966
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p0
|
Prémisse: {{premise}}
Hypothèse: {{hypothesis}}
La prémisse et l'hypothèse sont
|
2B
| 1,000
| 34.691
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p1
|
Quel est le rapport logique entre les deux phrases suivantes parmis :
A. L’hypothèse peut être inférée à partir de la prémisse.
B. L’hypothèse et la prémisse sont indépendantes.
C. L’hypothèse et la prémisse sont en contradiction.
Prémisse : {{premise}}
Hypothèse : {{hypothesis}}
Réponse :
|
2B
| 1,000
| 33.373
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p2
|
Quelle est la relation entre l'hypothèse suivante : {{hypothesis}} et la prémisse : {{premise}} ?
Réponse :
|
2B
| 1,000
| 33.333
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
2B
| 1,000
| 24.505
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
2B
| 1,000
| 24.441
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p2
|
Parmis les options suivantes :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Laquelle répond à la question : {{question.strip()}} ?
|
2B
| 1,000
| 25.189
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
2B
| 1,000
| 26.73
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
2B
| 1,000
| 22.912
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p2
|
Parmis les options suivantes :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Laquelle répond à la question : {{question.strip()}} ?
|
2B
| 1,000
| 26.014
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p0
|
Avis Client: {{text}}
A propos du thème "{{topic}}", l'avis client est
|
2B
| 1,000
| 34
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p1
|
Quelle est l’opinion au sujet de {{topic}} dans l’avis suivant ?
{{text}}
Est-t elle positive, neutre ou négative ?
|
2B
| 1,000
| 33.333
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p2
|
Avis client : {{text}}
Quel est l’avis du client sur {{topic}} ?
|
2B
| 1,000
| 33.333
|
Top
|
Reading comprehension
|
fquad
|
fquad_p1
|
{{context}}
Question : {{question}}
En prenant en compte le texte précédent, la réponse à la question est
|
4B
| 2,000
| 1.734
|
Top
|
Reading comprehension
|
fquad
|
fquad_p2
|
{{context}}
En répondant en référençant le texte précédent, {{question}}
|
4B
| 2,000
| 0
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p0
|
Texte : {{flores_passage}}
Question : {{question}}
A : {{mc_answer1}}
B : {{mc_answer2}}
C : {{mc_answer3}}
D : {{mc_answer4}}
Réponse :
|
4B
| 2,000
| 24.444
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p1
|
{{flores_passage}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Réponse :
|
4B
| 2,000
| 26.444
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Texte : {{flores_passage}}
Réponse :
|
4B
| 2,000
| 27.222
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p0
|
La phrase suivante est correcte grammaticalement:
|
4B
| 2,000
| 68.908
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
4B
| 2,000
| 70.588
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
4B
| 2,000
| 24.37
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p0
|
Context: {{context}}
|
4B
| 2,000
| 43.662
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p1
|
{{context}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Réponse :
|
4B
| 2,000
| 16.901
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Texte : {{context}}
Réponse :
|
4B
| 2,000
| 16.901
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p0
|
La phrase suivante est logique sémantiquement:
|
4B
| 2,000
| 65.546
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
4B
| 2,000
| 53.782
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
4B
| 2,000
| 15.966
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p0
|
Prémisse: {{premise}}
Hypothèse: {{hypothesis}}
La prémisse et l'hypothèse sont
|
4B
| 2,000
| 33.333
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p1
|
Quel est le rapport logique entre les deux phrases suivantes parmis :
A. L’hypothèse peut être inférée à partir de la prémisse.
B. L’hypothèse et la prémisse sont indépendantes.
C. L’hypothèse et la prémisse sont en contradiction.
Prémisse : {{premise}}
Hypothèse : {{hypothesis}}
Réponse :
|
4B
| 2,000
| 33.553
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p2
|
Quelle est la relation entre l'hypothèse suivante : {{hypothesis}} et la prémisse : {{premise}} ?
Réponse :
|
4B
| 2,000
| 33.513
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
4B
| 2,000
| 25.552
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
4B
| 2,000
| 25.303
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p2
|
Parmis les options suivantes :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Laquelle répond à la question : {{question.strip()}} ?
|
4B
| 2,000
| 25.474
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
4B
| 2,000
| 25.776
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
4B
| 2,000
| 21.957
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p2
|
Parmis les options suivantes :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Laquelle répond à la question : {{question.strip()}} ?
|
4B
| 2,000
| 22.912
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p0
|
Avis Client: {{text}}
A propos du thème "{{topic}}", l'avis client est
|
4B
| 2,000
| 35.833
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p1
|
Quelle est l’opinion au sujet de {{topic}} dans l’avis suivant ?
{{text}}
Est-t elle positive, neutre ou négative ?
|
4B
| 2,000
| 42.833
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p2
|
Avis client : {{text}}
Quel est l’avis du client sur {{topic}} ?
|
4B
| 2,000
| 35.5
|
Top
|
Reading comprehension
|
fquad
|
fquad_p1
|
{{context}}
Question : {{question}}
En prenant en compte le texte précédent, la réponse à la question est
|
6B
| 3,000
| 2.379
|
Top
|
Reading comprehension
|
fquad
|
fquad_p2
|
{{context}}
En répondant en référençant le texte précédent, {{question}}
|
6B
| 3,000
| 0.433
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p0
|
Texte : {{flores_passage}}
Question : {{question}}
A : {{mc_answer1}}
B : {{mc_answer2}}
C : {{mc_answer3}}
D : {{mc_answer4}}
Réponse :
|
6B
| 3,000
| 26.778
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p1
|
{{flores_passage}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Réponse :
|
6B
| 3,000
| 26.111
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Texte : {{flores_passage}}
Réponse :
|
6B
| 3,000
| 26.889
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p0
|
La phrase suivante est correcte grammaticalement:
|
6B
| 3,000
| 68.908
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
6B
| 3,000
| 72.269
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
6B
| 3,000
| 24.37
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p0
|
Context: {{context}}
|
6B
| 3,000
| 46.479
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p1
|
{{context}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Réponse :
|
6B
| 3,000
| 16.901
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Texte : {{context}}
Réponse :
|
6B
| 3,000
| 18.31
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p0
|
La phrase suivante est logique sémantiquement:
|
6B
| 3,000
| 68.908
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
6B
| 3,000
| 56.303
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
6B
| 3,000
| 16.807
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p0
|
Prémisse: {{premise}}
Hypothèse: {{hypothesis}}
La prémisse et l'hypothèse sont
|
6B
| 3,000
| 33.333
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p1
|
Quel est le rapport logique entre les deux phrases suivantes parmis :
A. L’hypothèse peut être inférée à partir de la prémisse.
B. L’hypothèse et la prémisse sont indépendantes.
C. L’hypothèse et la prémisse sont en contradiction.
Prémisse : {{premise}}
Hypothèse : {{hypothesis}}
Réponse :
|
6B
| 3,000
| 33.673
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p2
|
Quelle est la relation entre l'hypothèse suivante : {{hypothesis}} et la prémisse : {{premise}} ?
Réponse :
|
6B
| 3,000
| 33.353
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
6B
| 3,000
| 25.431
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
6B
| 3,000
| 25.345
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p2
|
Parmis les options suivantes :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Laquelle répond à la question : {{question.strip()}} ?
|
6B
| 3,000
| 25.531
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
6B
| 3,000
| 26.73
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
6B
| 3,000
| 26.73
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p2
|
Parmis les options suivantes :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Laquelle répond à la question : {{question.strip()}} ?
|
6B
| 3,000
| 26.492
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p0
|
Avis Client: {{text}}
A propos du thème "{{topic}}", l'avis client est
|
6B
| 3,000
| 42.167
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p1
|
Quelle est l’opinion au sujet de {{topic}} dans l’avis suivant ?
{{text}}
Est-t elle positive, neutre ou négative ?
|
6B
| 3,000
| 39.667
|
Top
|
Entailment
|
topic_based_nli
|
topic_based_nli_p2
|
Avis client : {{text}}
Quel est l’avis du client sur {{topic}} ?
|
6B
| 3,000
| 42.667
|
Top
|
Reading comprehension
|
fquad
|
fquad_p1
|
{{context}}
Question : {{question}}
En prenant en compte le texte précédent, la réponse à la question est
|
8B
| 4,000
| 1.649
|
Top
|
Reading comprehension
|
fquad
|
fquad_p2
|
{{context}}
En répondant en référençant le texte précédent, {{question}}
|
8B
| 4,000
| 0.685
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p0
|
Texte : {{flores_passage}}
Question : {{question}}
A : {{mc_answer1}}
B : {{mc_answer2}}
C : {{mc_answer3}}
D : {{mc_answer4}}
Réponse :
|
8B
| 4,000
| 24.778
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p1
|
{{flores_passage}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Réponse :
|
8B
| 4,000
| 25.889
|
Top
|
Reading comprehension
|
frabelebele
|
frabelebele_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{mc_answer1}}
B. {{mc_answer2}}
C. {{mc_answer3}}
D. {{mc_answer4}}
Texte : {{flores_passage}}
Réponse :
|
8B
| 4,000
| 26.889
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p0
|
La phrase suivante est correcte grammaticalement:
|
8B
| 4,000
| 74.79
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
8B
| 4,000
| 67.227
|
Top
|
Language-specific & world knowledge
|
french_bench_grammar
|
french_bench_grammar_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
8B
| 4,000
| 26.05
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p0
|
Context: {{context}}
|
8B
| 4,000
| 52.113
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p1
|
{{context}}
En se basant sur le texte qui précède, laquelle des options ci-dessous répond à la question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Réponse :
|
8B
| 4,000
| 23.944
|
Top
|
Reading comprehension
|
french_bench_reading
|
french_bench_reading_p2
|
Répondez à la question suivante avec l’une des options proposées en vous basant sur le texte ci-dessous.
Question : {{question}}
Options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
Texte : {{context}}
Réponse :
|
8B
| 4,000
| 21.127
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p0
|
La phrase suivante est logique sémantiquement:
|
8B
| 4,000
| 70.588
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p1
|
Remplissez les trous dans la phrase : « {{question.strip()}} » avec une des options suivantes :
- {{answerA}}
- {{answerB}}
- {{answerC}}
- {{answerD}}
|
8B
| 4,000
| 57.143
|
Top
|
Language-specific & world knowledge
|
french_bench_vocabulary
|
french_bench_vocabulary_p2
|
Laquelle de ces options :
A. {{answerA}}
B. {{answerB}}
C. {{answerC}}
D. {{answerD}}
remplie au mieux la phrase suivante : {{question.strip()}} ?
|
8B
| 4,000
| 20.168
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p0
|
Prémisse: {{premise}}
Hypothèse: {{hypothesis}}
La prémisse et l'hypothèse sont
|
8B
| 4,000
| 33.333
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p1
|
Quel est le rapport logique entre les deux phrases suivantes parmis :
A. L’hypothèse peut être inférée à partir de la prémisse.
B. L’hypothèse et la prémisse sont indépendantes.
C. L’hypothèse et la prémisse sont en contradiction.
Prémisse : {{premise}}
Hypothèse : {{hypothesis}}
Réponse :
|
8B
| 4,000
| 33.333
|
Top
|
Entailment
|
french_xnli
|
french_xnli_p2
|
Quelle est la relation entre l'hypothèse suivante : {{hypothesis}} et la prémisse : {{premise}} ?
Réponse :
|
8B
| 4,000
| 33.253
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
8B
| 4,000
| 25.509
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
8B
| 4,000
| 25.21
|
Top
|
Language-specific & world knowledge
|
global_mmlu_french
|
global_mmlu_french_p2
|
Parmis les options suivantes :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Laquelle répond à la question : {{question.strip()}} ?
|
8B
| 4,000
| 24.655
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p0
|
{{question.strip()}}
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
8B
| 4,000
| 24.344
|
Top
|
Language-specific & world knowledge
|
include_french
|
include_french_p1
|
Question à choix multiple : {{question.strip()}}
Choix :
A. {{option_a}}
B. {{option_b}}
C. {{option_c}}
D. {{option_d}}
Réponse :
|
8B
| 4,000
| 24.105
|
HPLT 3.0: Details on WDS-based Sampling Evaluation Results
Dataset Description
This dataset contains fine-grained results from our HPLT 3.0 release evaluations comparing the new HPLT 3.0 corpora sampled using different Web Document Scorer (WDS) thresholds, focusing on Spanish and French. We compare three configurations: Top, Random, and Bottom. Random sampling represents the default approach, drawing uniformly on the full corpus, while Top and Bottom take advantage of the sorting by WDS levels and sequentially draw 100B training tokens from either end of the corpus. We pretrain 2.2B Llama-style decoder models on 100B tokens for each selected language and evaluate them using HPLT-E, a multilingual evaluation framework for comprehensive multi-prompt k-shot evaluation across 124 tasks and 500+ prompts in nine typologically diverse languages: Spanish (spa_Latn), French (fra_Latn), Czech (ces_Latn), Ukrainian (ukr_Cyrl), Finnish (fin_Latn), Catalan (cat_Latn), Galician (glg_Latn), Basque (eus_Latn), and Norwegian (Bokmål and Nynorsk; nor_Latn).
- Curated by: High Performance Language Technologies (HPLT)
- Languages: Spanish and French
- Paper: arxiv.org/abs/2511.01066
- Repository: github.com/hplt-project/hplt-e
- License: Apache 2.0
Please find more details in our paper and GitHub repository.
Uses
This dataset is intended for reproducibility and research purposes. Find an example on how to access the results:
from datasets import load_dataset
dataset = load_dataset("HPLT/2508-wds-evals", "spa_Latn", split="results").to_pandas()
Dataset Structure
Dataset Instances
Each dataset instance looks as follows:
{
'corpus': 'Bottom',
'category': 'Paraphrase detection',
'dataset': 'paws_es',
'task': 'paws_es_p2',
'prompt': 'Oración 1: {{sentence1}}\nOración 2: {{sentence2}}\nPregunta: ¿Las oraciones 1 y 2 expresan el mismo significado? ¿Sí o no?\nRespuesta:',
'model': '50B',
'ckpt_num': 24000,
'score': 45.35
}
Dataset Fields
corpus: corpus name (Top,Random,Bottom)category: task categorydataset: evaluation dataset nametask: evaluation task (refers to a specific prompt)prompt: prompt used for evaluationmodel: number of pretraining tokens (B)ckpt_num: number identifier formodelscore: standard metric performance score
Cite Us
@article{oepen2025hplt,
title={HPLT\~{} 3.0: Very Large-Scale Multilingual Resources for LLM and MT. Mono-and Bi-lingual Data, Multilingual Evaluation, and Pre-Trained Models},
author={Oepen, Stephan and Arefev, Nikolay and Aulamo, Mikko and Ba{\~n}{\'o}n, Marta and Buljan, Maja and Burchell, Laurie and Charpentier, Lucas and Chen, Pinzhen and Fedorova, Mariya and de Gibert, Ona and others},
journal={arXiv preprint arXiv:2511.01066},
year={2025}
}
Contact Us
- Vladislav Mikhailov [email protected]
- Stephan Oepen [email protected]
- Downloads last month
- 26