data_source stringclasses 1
value | prompt listlengths 2 2 | ability stringclasses 1
value | reward_model dict | extra_info dict |
|---|---|---|---|---|
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sv_philosophy__0d1ce3e7__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1004.8859846228888, \"nll_base\": 6.9126293659210205, \"eval_ref\": \"global_mmlu_full_sv_philosophy__0d1ce3e7__s000__tb262144__v000... | {
"index": "global_mmlu_full_sv_philosophy__0d1ce3e7__s000__tb262144__v000",
"uid": "global_mmlu_full_sv_philosophy__0d1ce3e7__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_business_ethics__8406d22d__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1434.3987392042347, \"nll_base\": 7.268501043319702, \"eval_ref\": \"global_mmlu_full_sw_business_ethics__8406d22d__s000__tb262... | {
"index": "global_mmlu_full_sw_business_ethics__8406d22d__s000__tb262144__v000",
"uid": "global_mmlu_full_sw_business_ethics__8406d22d__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1090.3669930763838, \"nll_base\": 6.994269609451294, \"eval_ref\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000_... | {
"index": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb131072__v001",
"uid": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb131072__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb262144__v002\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1090.3669930763838, \"nll_base\": 6.994269609451294, \"eval_ref\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000_... | {
"index": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb262144__v002",
"uid": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb262144__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1090.3669930763838, \"nll_base\": 6.994269609451294, \"eval_ref\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__t... | {
"index": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb65536__v000",
"uid": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s001__tb65536__v003\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 938.5624388689503, \"nll_base\": 6.844349384307861, \"eval_ref\": \"global_mmlu_full_sw_clinical_knowledge__9ded7e44__s001__tb... | {
"index": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s001__tb65536__v003",
"uid": "global_mmlu_full_sw_clinical_knowledge__9ded7e44__s001__tb65536__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_computer_security__4dc3a9c8__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1076.2824082629456, \"nll_base\": 6.9812681674957275, \"eval_ref\": \"global_mmlu_full_sw_computer_security__4dc3a9c8__s000__... | {
"index": "global_mmlu_full_sw_computer_security__4dc3a9c8__s000__tb131072__v000",
"uid": "global_mmlu_full_sw_computer_security__4dc3a9c8__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_high_school_geography__e4b2bf6b__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1066.953627710169, \"nll_base\": 6.972562789916992, \"eval_ref\": \"global_mmlu_full_sw_high_school_geography__e4b2bf6b__... | {
"index": "global_mmlu_full_sw_high_school_geography__e4b2bf6b__s000__tb262144__v000",
"uid": "global_mmlu_full_sw_high_school_geography__e4b2bf6b__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_high_school_macroeconomics__b6cbaf17__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 906.2882432337535, \"nll_base\": 6.809357404708862, \"eval_ref\": \"global_mmlu_full_sw_high_school_macroeconomics__b6... | {
"index": "global_mmlu_full_sw_high_school_macroeconomics__b6cbaf17__s000__tb65536__v000",
"uid": "global_mmlu_full_sw_high_school_macroeconomics__b6cbaf17__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_high_school_us_history__86473553__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1203.4546035791043, \"nll_base\": 7.092951536178589, \"eval_ref\": \"global_mmlu_full_sw_high_school_us_history__86473553_... | {
"index": "global_mmlu_full_sw_high_school_us_history__86473553__s000__tb65536__v000",
"uid": "global_mmlu_full_sw_high_school_us_history__86473553__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_international_law__ec71ad5e__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 725.5014282205823, \"nll_base\": 6.586863040924072, \"eval_ref\": \"global_mmlu_full_sw_international_law__ec71ad5e__s000__tb65... | {
"index": "global_mmlu_full_sw_international_law__ec71ad5e__s000__tb65536__v000",
"uid": "global_mmlu_full_sw_international_law__ec71ad5e__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1262.1559397378294, \"nll_base\": 7.140576601028442, \"eval_ref\": \"global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb1... | {
"index": "global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb131072__v000",
"uid": "global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb131072__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb262144__v001\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1262.1559397378294, \"nll_base\": 7.140576601028442, \"eval_ref\": \"global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb2... | {
"index": "global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb262144__v001",
"uid": "global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb262144__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb65536__v002\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1262.1559397378294, \"nll_base\": 7.140576601028442, \"eval_ref\": \"global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb655... | {
"index": "global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb65536__v002",
"uid": "global_mmlu_full_sw_medical_genetics__9151a6d1__s000__tb65536__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_professional_medicine__2bdd120d__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1142.1840874198667, \"nll_base\": 7.0406975746154785, \"eval_ref\": \"global_mmlu_full_sw_professional_medicine__2bdd120d... | {
"index": "global_mmlu_full_sw_professional_medicine__2bdd120d__s000__tb262144__v000",
"uid": "global_mmlu_full_sw_professional_medicine__2bdd120d__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_public_relations__41bf7917__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1162.6549602161008, \"nll_base\": 7.058461427688599, \"eval_ref\": \"global_mmlu_full_sw_public_relations__41bf7917__s000__tb655... | {
"index": "global_mmlu_full_sw_public_relations__41bf7917__s000__tb65536__v000",
"uid": "global_mmlu_full_sw_public_relations__41bf7917__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_sw_us_foreign_policy__c4136c6f__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 963.0956307155544, \"nll_base\": 6.870152711868286, \"eval_ref\": \"global_mmlu_full_sw_us_foreign_policy__c4136c6f__s000__tb... | {
"index": "global_mmlu_full_sw_us_foreign_policy__c4136c6f__s000__tb131072__v000",
"uid": "global_mmlu_full_sw_us_foreign_policy__c4136c6f__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_clinical_knowledge__5317a043__s001__tb131072__v003\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 880.2800177971417, \"nll_base\": 6.780240058898926, \"eval_ref\": \"global_mmlu_full_tr_clinical_knowledge__5317a043__s001__... | {
"index": "global_mmlu_full_tr_clinical_knowledge__5317a043__s001__tb131072__v003",
"uid": "global_mmlu_full_tr_clinical_knowledge__5317a043__s001__tb131072__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_college_chemistry__028c5a23__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1073.4094428086273, \"nll_base\": 6.97859525680542, \"eval_ref\": \"global_mmlu_full_tr_college_chemistry__028c5a23__s000__tb65... | {
"index": "global_mmlu_full_tr_college_chemistry__028c5a23__s000__tb65536__v000",
"uid": "global_mmlu_full_tr_college_chemistry__028c5a23__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_college_mathematics__9b1a71a4__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 902.500374828859, \"nll_base\": 6.805169105529785, \"eval_ref\": \"global_mmlu_full_tr_college_mathematics__9b1a71a4__s000__t... | {
"index": "global_mmlu_full_tr_college_mathematics__9b1a71a4__s000__tb65536__v000",
"uid": "global_mmlu_full_tr_college_mathematics__9b1a71a4__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_high_school_mathematics__61f37793__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1470.8293791406484, \"nll_base\": 7.29358172416687, \"eval_ref\": \"global_mmlu_full_tr_high_school_mathematics__61f37793... | {
"index": "global_mmlu_full_tr_high_school_mathematics__61f37793__s000__tb65536__v000",
"uid": "global_mmlu_full_tr_high_school_mathematics__61f37793__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_jurisprudence__d3604751__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 902.1968169902233, \"nll_base\": 6.804832696914673, \"eval_ref\": \"global_mmlu_full_tr_jurisprudence__d3604751__s000__tb131072__... | {
"index": "global_mmlu_full_tr_jurisprudence__d3604751__s000__tb131072__v000",
"uid": "global_mmlu_full_tr_jurisprudence__d3604751__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_management__f83ea1be__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1290.9127077532576, \"nll_base\": 7.163104772567749, \"eval_ref\": \"global_mmlu_full_tr_management__f83ea1be__s000__tb131072__v000\... | {
"index": "global_mmlu_full_tr_management__f83ea1be__s000__tb131072__v000",
"uid": "global_mmlu_full_tr_management__f83ea1be__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 998.5877487560696, \"nll_base\": 6.906342029571533, \"eval_ref\": \"global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb13... | {
"index": "global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb131072__v001",
"uid": "global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb131072__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb262144__v002\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 998.5877487560696, \"nll_base\": 6.906342029571533, \"eval_ref\": \"global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb26... | {
"index": "global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb262144__v002",
"uid": "global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb262144__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 998.5877487560696, \"nll_base\": 6.906342029571533, \"eval_ref\": \"global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb6553... | {
"index": "global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb65536__v000",
"uid": "global_mmlu_full_tr_medical_genetics__f6f70901__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_moral_disputes__1ecace44__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 820.4085018055015, \"nll_base\": 6.7098023891448975, \"eval_ref\": \"global_mmlu_full_tr_moral_disputes__1ecace44__s000__tb13107... | {
"index": "global_mmlu_full_tr_moral_disputes__1ecace44__s000__tb131072__v000",
"uid": "global_mmlu_full_tr_moral_disputes__1ecace44__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_tr_moral_scenarios__7716e581__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 814.0496509515655, \"nll_base\": 6.702021360397339, \"eval_ref\": \"global_mmlu_full_tr_moral_scenarios__7716e581__s000__tb1310... | {
"index": "global_mmlu_full_tr_moral_scenarios__7716e581__s000__tb131072__v000",
"uid": "global_mmlu_full_tr_moral_scenarios__7716e581__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_uk_business_ethics__4ae28985__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1543.7233889987297, \"nll_base\": 7.341952562332153, \"eval_ref\": \"global_mmlu_full_uk_business_ethics__4ae28985__s000__tb65536... | {
"index": "global_mmlu_full_uk_business_ethics__4ae28985__s000__tb65536__v000",
"uid": "global_mmlu_full_uk_business_ethics__4ae28985__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_uk_electrical_engineering__3f230420__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 879.8767302420988, \"nll_base\": 6.779781818389893, \"eval_ref\": \"global_mmlu_full_uk_electrical_engineering__3f230420... | {
"index": "global_mmlu_full_uk_electrical_engineering__3f230420__s000__tb131072__v000",
"uid": "global_mmlu_full_uk_electrical_engineering__3f230420__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_uk_high_school_chemistry__336efa5b__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1440.0210569795538, \"nll_base\": 7.272413015365601, \"eval_ref\": \"global_mmlu_full_uk_high_school_chemistry__336efa5b_... | {
"index": "global_mmlu_full_uk_high_school_chemistry__336efa5b__s000__tb131072__v000",
"uid": "global_mmlu_full_uk_high_school_chemistry__336efa5b__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_uk_moral_scenarios__7d58480f__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 898.727480879595, \"nll_base\": 6.800979852676392, \"eval_ref\": \"global_mmlu_full_uk_moral_scenarios__7d58480f__s000__tb13107... | {
"index": "global_mmlu_full_uk_moral_scenarios__7d58480f__s000__tb131072__v000",
"uid": "global_mmlu_full_uk_moral_scenarios__7d58480f__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_uk_prehistory__a1009846__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1463.047911372302, \"nll_base\": 7.2882771492004395, \"eval_ref\": \"global_mmlu_full_uk_prehistory__a1009846__s000__tb65536__v000\", ... | {
"index": "global_mmlu_full_uk_prehistory__a1009846__s000__tb65536__v000",
"uid": "global_mmlu_full_uk_prehistory__a1009846__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_uk_professional_accounting__9a5d7a26__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1071.3801204605338, \"nll_base\": 6.976702928543091, \"eval_ref\": \"global_mmlu_full_uk_professional_accounting__9a5d7a2... | {
"index": "global_mmlu_full_uk_professional_accounting__9a5d7a26__s000__tb65536__v000",
"uid": "global_mmlu_full_uk_professional_accounting__9a5d7a26__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_abstract_algebra__a9509d1f__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 506.6130211909053, \"nll_base\": 6.227747440338135, \"eval_ref\": \"global_mmlu_full_vi_abstract_algebra__a9509d1f__s000__tb26... | {
"index": "global_mmlu_full_vi_abstract_algebra__a9509d1f__s000__tb262144__v000",
"uid": "global_mmlu_full_vi_abstract_algebra__a9509d1f__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 760.2033246016803, \"nll_base\": 6.6335859298706055, \"eval_ref\": \"global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000_... | {
"index": "global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000__tb131072__v001",
"uid": "global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000__tb131072__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 760.2033246016803, \"nll_base\": 6.6335859298706055, \"eval_ref\": \"global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000__t... | {
"index": "global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000__tb65536__v000",
"uid": "global_mmlu_full_vi_clinical_knowledge__9cdf9b22__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_college_chemistry__d47b16a8__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 662.7238546661456, \"nll_base\": 6.496358394622803, \"eval_ref\": \"global_mmlu_full_vi_college_chemistry__d47b16a8__s000__tb65... | {
"index": "global_mmlu_full_vi_college_chemistry__d47b16a8__s000__tb65536__v000",
"uid": "global_mmlu_full_vi_college_chemistry__d47b16a8__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_college_computer_science__0c20809f__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 713.765853080827, \"nll_base\": 6.570554971694946, \"eval_ref\": \"global_mmlu_full_vi_college_computer_science__0c208... | {
"index": "global_mmlu_full_vi_college_computer_science__0c20809f__s000__tb262144__v000",
"uid": "global_mmlu_full_vi_college_computer_science__0c20809f__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_elementary_mathematics__216cfd8c__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1082.8057698699163, \"nll_base\": 6.987310886383057, \"eval_ref\": \"global_mmlu_full_vi_elementary_mathematics__216cfd8... | {
"index": "global_mmlu_full_vi_elementary_mathematics__216cfd8c__s000__tb262144__v000",
"uid": "global_mmlu_full_vi_elementary_mathematics__216cfd8c__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_high_school_computer_science__61555b11__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 711.4214662551266, \"nll_base\": 6.567265033721924, \"eval_ref\": \"global_mmlu_full_vi_high_school_computer_scien... | {
"index": "global_mmlu_full_vi_high_school_computer_science__61555b11__s000__tb131072__v000",
"uid": "global_mmlu_full_vi_high_school_computer_science__61555b11__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_marketing__2bf34c14__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 700.9612197933678, \"nll_base\": 6.552452564239502, \"eval_ref\": \"global_mmlu_full_vi_marketing__2bf34c14__s000__tb131072__v000\", ... | {
"index": "global_mmlu_full_vi_marketing__2bf34c14__s000__tb131072__v000",
"uid": "global_mmlu_full_vi_marketing__2bf34c14__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb131072__v002\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 890.8323224619678, \"nll_base\": 6.792156219482422, \"eval_ref\": \"global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb13... | {
"index": "global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb131072__v002",
"uid": "global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb131072__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb65536__v001\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 890.8323224619678, \"nll_base\": 6.792156219482422, \"eval_ref\": \"global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb6553... | {
"index": "global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb65536__v001",
"uid": "global_mmlu_full_vi_medical_genetics__2cb206bf__s000__tb65536__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_medical_genetics__2cb206bf__s001__tb262144__v003\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 668.009093469798, \"nll_base\": 6.5043017864227295, \"eval_ref\": \"global_mmlu_full_vi_medical_genetics__2cb206bf__s001__tb26... | {
"index": "global_mmlu_full_vi_medical_genetics__2cb206bf__s001__tb262144__v003",
"uid": "global_mmlu_full_vi_medical_genetics__2cb206bf__s001__tb262144__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_miscellaneous__9f4f99c8__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 983.3428438554834, \"nll_base\": 6.890957832336426, \"eval_ref\": \"global_mmlu_full_vi_miscellaneous__9f4f99c8__s000__tb131072__... | {
"index": "global_mmlu_full_vi_miscellaneous__9f4f99c8__s000__tb131072__v000",
"uid": "global_mmlu_full_vi_miscellaneous__9f4f99c8__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_vi_professional_medicine__f504c5b9__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 477.1741357815496, \"nll_base\": 6.167881488800049, \"eval_ref\": \"global_mmlu_full_vi_professional_medicine__f504c5b9__... | {
"index": "global_mmlu_full_vi_professional_medicine__f504c5b9__s000__tb131072__v000",
"uid": "global_mmlu_full_vi_professional_medicine__f504c5b9__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb131072__v002\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 805.1450085875967, \"nll_base\": 6.6910223960876465, \"eval_ref\": \"global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000_... | {
"index": "global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb131072__v002",
"uid": "global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb131072__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 805.1450085875967, \"nll_base\": 6.6910223960876465, \"eval_ref\": \"global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000_... | {
"index": "global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb262144__v000",
"uid": "global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb262144__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb65536__v001\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 805.1450085875967, \"nll_base\": 6.6910223960876465, \"eval_ref\": \"global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__t... | {
"index": "global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb65536__v001",
"uid": "global_mmlu_full_yo_clinical_knowledge__e7f72f79__s000__tb65536__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_college_biology__d25b7c91__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 691.2125540865879, \"nll_base\": 6.538447380065918, \"eval_ref\": \"global_mmlu_full_yo_college_biology__d25b7c91__s000__tb2621... | {
"index": "global_mmlu_full_yo_college_biology__d25b7c91__s000__tb262144__v000",
"uid": "global_mmlu_full_yo_college_biology__d25b7c91__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_computer_security__1b333ebc__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 940.5186644246924, \"nll_base\": 6.846431493759155, \"eval_ref\": \"global_mmlu_full_yo_computer_security__1b333ebc__s000__tb65... | {
"index": "global_mmlu_full_yo_computer_security__1b333ebc__s000__tb65536__v000",
"uid": "global_mmlu_full_yo_computer_security__1b333ebc__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_high_school_microeconomics__55ffb008__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 438.6131975234879, \"nll_base\": 6.083617925643921, \"eval_ref\": \"global_mmlu_full_yo_high_school_microeconomics__... | {
"index": "global_mmlu_full_yo_high_school_microeconomics__55ffb008__s000__tb131072__v000",
"uid": "global_mmlu_full_yo_high_school_microeconomics__55ffb008__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_high_school_physics__6ba08ee5__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 970.3938283989803, \"nll_base\": 6.877701997756958, \"eval_ref\": \"global_mmlu_full_yo_high_school_physics__6ba08ee5__s000__... | {
"index": "global_mmlu_full_yo_high_school_physics__6ba08ee5__s000__tb65536__v000",
"uid": "global_mmlu_full_yo_high_school_physics__6ba08ee5__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_high_school_world_history__b0e0d051__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 372.5820802194632, \"nll_base\": 5.920457363128662, \"eval_ref\": \"global_mmlu_full_yo_high_school_world_history__b0... | {
"index": "global_mmlu_full_yo_high_school_world_history__b0e0d051__s000__tb262144__v000",
"uid": "global_mmlu_full_yo_high_school_world_history__b0e0d051__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_jurisprudence__c520a272__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 899.9183432309428, \"nll_base\": 6.802304029464722, \"eval_ref\": \"global_mmlu_full_yo_jurisprudence__c520a272__s000__tb262144__... | {
"index": "global_mmlu_full_yo_jurisprudence__c520a272__s000__tb262144__v000",
"uid": "global_mmlu_full_yo_jurisprudence__c520a272__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_yo_professional_psychology__bd5c98c1__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 551.435454473906, \"nll_base\": 6.312524795532227, \"eval_ref\": \"global_mmlu_full_yo_professional_psychology__bd5c98c... | {
"index": "global_mmlu_full_yo_professional_psychology__bd5c98c1__s000__tb131072__v000",
"uid": "global_mmlu_full_yo_professional_psychology__bd5c98c1__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_zh_college_biology__baead91a__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1002.0227363609172, \"nll_base\": 6.909775972366333, \"eval_ref\": \"global_mmlu_full_zh_college_biology__baead91a__s000__tb131... | {
"index": "global_mmlu_full_zh_college_biology__baead91a__s000__tb131072__v000",
"uid": "global_mmlu_full_zh_college_biology__baead91a__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_zh_high_school_computer_science__1d87827d__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1456.007623274032, \"nll_base\": 7.283453464508057, \"eval_ref\": \"global_mmlu_full_zh_high_school_computer_scien... | {
"index": "global_mmlu_full_zh_high_school_computer_science__1d87827d__s000__tb131072__v000",
"uid": "global_mmlu_full_zh_high_school_computer_science__1d87827d__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_zh_high_school_us_history__d7ffd2f0__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 609.0389129231348, \"nll_base\": 6.411882162094116, \"eval_ref\": \"global_mmlu_full_zh_high_school_us_history__d7ffd2f0__... | {
"index": "global_mmlu_full_zh_high_school_us_history__d7ffd2f0__s000__tb65536__v000",
"uid": "global_mmlu_full_zh_high_school_us_history__d7ffd2f0__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_zh_machine_learning__fba7ef27__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 809.3385414071857, \"nll_base\": 6.69621729850769, \"eval_ref\": \"global_mmlu_full_zh_machine_learning__fba7ef27__s000__tb131... | {
"index": "global_mmlu_full_zh_machine_learning__fba7ef27__s000__tb131072__v000",
"uid": "global_mmlu_full_zh_machine_learning__fba7ef27__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_zh_philosophy__00a96acf__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 900.6414769985726, \"nll_base\": 6.803107261657715, \"eval_ref\": \"global_mmlu_full_zh_philosophy__00a96acf__s000__tb65536__v000\", \... | {
"index": "global_mmlu_full_zh_philosophy__00a96acf__s000__tb65536__v000",
"uid": "global_mmlu_full_zh_philosophy__00a96acf__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_full_zh_sociology__00e0fdbc__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 865.1168544791146, \"nll_base\": 6.762864589691162, \"eval_ref\": \"global_mmlu_full_zh_sociology__00e0fdbc__s000__tb131072__v000\", ... | {
"index": "global_mmlu_full_zh_sociology__00e0fdbc__s000__tb131072__v000",
"uid": "global_mmlu_full_zh_sociology__00e0fdbc__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_hi_medical__eeb9681b__s000__tb131072__v002\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 568.3469278799671, \"nll_base\": 6.342732020786831, \"eval_ref\": \"global_mmlu_hi_medical__eeb9681b__s000__tb131072__v002\", \"oracle_delta... | {
"index": "global_mmlu_hi_medical__eeb9681b__s000__tb131072__v002",
"uid": "global_mmlu_hi_medical__eeb9681b__s000__tb131072__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_hi_medical__eeb9681b__s000__tb65536__v001\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 568.3469278799671, \"nll_base\": 6.342732020786831, \"eval_ref\": \"global_mmlu_hi_medical__eeb9681b__s000__tb65536__v001\", \"oracle_delta\":... | {
"index": "global_mmlu_hi_medical__eeb9681b__s000__tb65536__v001",
"uid": "global_mmlu_hi_medical__eeb9681b__s000__tb65536__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_id_medical__763a0fef__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1139.5673134828553, \"nll_base\": 7.038403919764927, \"eval_ref\": \"global_mmlu_id_medical__763a0fef__s000__tb131072__v000\", \"oracle_delt... | {
"index": "global_mmlu_id_medical__763a0fef__s000__tb131072__v000",
"uid": "global_mmlu_id_medical__763a0fef__s000__tb131072__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_id_medical__763a0fef__s000__tb262144__v001\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1139.5673134828553, \"nll_base\": 7.038403919764927, \"eval_ref\": \"global_mmlu_id_medical__763a0fef__s000__tb262144__v001\", \"oracle_delt... | {
"index": "global_mmlu_id_medical__763a0fef__s000__tb262144__v001",
"uid": "global_mmlu_id_medical__763a0fef__s000__tb262144__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_id_medical__763a0fef__s000__tb65536__v002\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1139.5673134828553, \"nll_base\": 7.038403919764927, \"eval_ref\": \"global_mmlu_id_medical__763a0fef__s000__tb65536__v002\", \"oracle_delta\"... | {
"index": "global_mmlu_id_medical__763a0fef__s000__tb65536__v002",
"uid": "global_mmlu_id_medical__763a0fef__s000__tb65536__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_id_medical__763a0fef__s001__tb131072__v003\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1199.5160359050353, \"nll_base\": 7.089673451014927, \"eval_ref\": \"global_mmlu_id_medical__763a0fef__s001__tb131072__v003\", \"oracle_delt... | {
"index": "global_mmlu_id_medical__763a0fef__s001__tb131072__v003",
"uid": "global_mmlu_id_medical__763a0fef__s001__tb131072__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_id_stem__239771cc__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1625.1987749170084, \"nll_base\": 7.393385410308838, \"eval_ref\": \"global_mmlu_id_stem__239771cc__s000__tb262144__v000\", \"oracle_delta\": 0... | {
"index": "global_mmlu_id_stem__239771cc__s000__tb262144__v000",
"uid": "global_mmlu_id_stem__239771cc__s000__tb262144__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_it_medical__f49754bd__s000__tb131072__v002\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1130.289751397519, \"nll_base\": 7.0302292960030695, \"eval_ref\": \"global_mmlu_it_medical__f49754bd__s000__tb131072__v002\", \"oracle_delt... | {
"index": "global_mmlu_it_medical__f49754bd__s000__tb131072__v002",
"uid": "global_mmlu_it_medical__f49754bd__s000__tb131072__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_it_medical__f49754bd__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1130.289751397519, \"nll_base\": 7.0302292960030695, \"eval_ref\": \"global_mmlu_it_medical__f49754bd__s000__tb262144__v000\", \"oracle_delt... | {
"index": "global_mmlu_it_medical__f49754bd__s000__tb262144__v000",
"uid": "global_mmlu_it_medical__f49754bd__s000__tb262144__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_it_medical__f49754bd__s000__tb65536__v001\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1130.289751397519, \"nll_base\": 7.0302292960030695, \"eval_ref\": \"global_mmlu_it_medical__f49754bd__s000__tb65536__v001\", \"oracle_delta\"... | {
"index": "global_mmlu_it_medical__f49754bd__s000__tb65536__v001",
"uid": "global_mmlu_it_medical__f49754bd__s000__tb65536__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_it_medical__f49754bd__s001__tb262144__v003\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1126.349876551077, \"nll_base\": 7.0267374856131415, \"eval_ref\": \"global_mmlu_it_medical__f49754bd__s001__tb262144__v003\", \"oracle_delt... | {
"index": "global_mmlu_it_medical__f49754bd__s001__tb262144__v003",
"uid": "global_mmlu_it_medical__f49754bd__s001__tb262144__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_ja_medical__842324eb__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 694.6264352520337, \"nll_base\": 6.543374197823661, \"eval_ref\": \"global_mmlu_ja_medical__842324eb__s000__tb131072__v001\", \"oracle_delta... | {
"index": "global_mmlu_ja_medical__842324eb__s000__tb131072__v001",
"uid": "global_mmlu_ja_medical__842324eb__s000__tb131072__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_ja_medical__842324eb__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 694.6264352520337, \"nll_base\": 6.543374197823661, \"eval_ref\": \"global_mmlu_ja_medical__842324eb__s000__tb65536__v000\", \"oracle_delta\":... | {
"index": "global_mmlu_ja_medical__842324eb__s000__tb65536__v000",
"uid": "global_mmlu_ja_medical__842324eb__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_ja_medical__842324eb__s001__tb65536__v003\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 708.9150405893696, \"nll_base\": 6.563735689435687, \"eval_ref\": \"global_mmlu_ja_medical__842324eb__s001__tb65536__v003\", \"oracle_delta\":... | {
"index": "global_mmlu_ja_medical__842324eb__s001__tb65536__v003",
"uid": "global_mmlu_ja_medical__842324eb__s001__tb65536__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_ja_social_sciences__47a38307__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 543.4184753163755, \"nll_base\": 6.297879695892334, \"eval_ref\": \"global_mmlu_ja_social_sciences__47a38307__s000__tb65536__v000\", \... | {
"index": "global_mmlu_ja_social_sciences__47a38307__s000__tb65536__v000",
"uid": "global_mmlu_ja_social_sciences__47a38307__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_ko_medical__57ef2659__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 863.7250135986154, \"nll_base\": 6.761254446847098, \"eval_ref\": \"global_mmlu_ko_medical__57ef2659__s000__tb65536__v000\", \"oracle_delta\":... | {
"index": "global_mmlu_ko_medical__57ef2659__s000__tb65536__v000",
"uid": "global_mmlu_ko_medical__57ef2659__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_pt_humanities__94f07495__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 551.7668646149539, \"nll_base\": 6.3131256103515625, \"eval_ref\": \"global_mmlu_pt_humanities__94f07495__s000__tb65536__v000\", \"oracle_d... | {
"index": "global_mmlu_pt_humanities__94f07495__s000__tb65536__v000",
"uid": "global_mmlu_pt_humanities__94f07495__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_pt_medical__7e8356b6__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 760.542175879832, \"nll_base\": 6.634031568254743, \"eval_ref\": \"global_mmlu_pt_medical__7e8356b6__s000__tb131072__v001\", \"oracle_delta\... | {
"index": "global_mmlu_pt_medical__7e8356b6__s000__tb131072__v001",
"uid": "global_mmlu_pt_medical__7e8356b6__s000__tb131072__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_pt_medical__7e8356b6__s000__tb262144__v002\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 760.542175879832, \"nll_base\": 6.634031568254743, \"eval_ref\": \"global_mmlu_pt_medical__7e8356b6__s000__tb262144__v002\", \"oracle_delta\... | {
"index": "global_mmlu_pt_medical__7e8356b6__s000__tb262144__v002",
"uid": "global_mmlu_pt_medical__7e8356b6__s000__tb262144__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_pt_medical__7e8356b6__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 760.542175879832, \"nll_base\": 6.634031568254743, \"eval_ref\": \"global_mmlu_pt_medical__7e8356b6__s000__tb65536__v000\", \"oracle_delta\": ... | {
"index": "global_mmlu_pt_medical__7e8356b6__s000__tb65536__v000",
"uid": "global_mmlu_pt_medical__7e8356b6__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_sw_medical__a90550ec__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 921.8869066252137, \"nll_base\": 6.826422555106027, \"eval_ref\": \"global_mmlu_sw_medical__a90550ec__s000__tb131072__v000\", \"oracle_delta... | {
"index": "global_mmlu_sw_medical__a90550ec__s000__tb131072__v000",
"uid": "global_mmlu_sw_medical__a90550ec__s000__tb131072__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_sw_medical__a90550ec__s000__tb262144__v001\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 921.8869066252137, \"nll_base\": 6.826422555106027, \"eval_ref\": \"global_mmlu_sw_medical__a90550ec__s000__tb262144__v001\", \"oracle_delta... | {
"index": "global_mmlu_sw_medical__a90550ec__s000__tb262144__v001",
"uid": "global_mmlu_sw_medical__a90550ec__s000__tb262144__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_sw_medical__a90550ec__s000__tb65536__v002\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 921.8869066252137, \"nll_base\": 6.826422555106027, \"eval_ref\": \"global_mmlu_sw_medical__a90550ec__s000__tb65536__v002\", \"oracle_delta\":... | {
"index": "global_mmlu_sw_medical__a90550ec__s000__tb65536__v002",
"uid": "global_mmlu_sw_medical__a90550ec__s000__tb65536__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_sw_medical__a90550ec__s001__tb131072__v003\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 894.1812715348434, \"nll_base\": 6.795908519199917, \"eval_ref\": \"global_mmlu_sw_medical__a90550ec__s001__tb131072__v003\", \"oracle_delta... | {
"index": "global_mmlu_sw_medical__a90550ec__s001__tb131072__v003",
"uid": "global_mmlu_sw_medical__a90550ec__s001__tb131072__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_yo_medical__bfb0ac36__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 686.5803730750644, \"nll_base\": 6.531723294939313, \"eval_ref\": \"global_mmlu_yo_medical__bfb0ac36__s000__tb131072__v001\", \"oracle_delta... | {
"index": "global_mmlu_yo_medical__bfb0ac36__s000__tb131072__v001",
"uid": "global_mmlu_yo_medical__bfb0ac36__s000__tb131072__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_yo_medical__bfb0ac36__s000__tb262144__v002\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 686.5803730750644, \"nll_base\": 6.531723294939313, \"eval_ref\": \"global_mmlu_yo_medical__bfb0ac36__s000__tb262144__v002\", \"oracle_delta... | {
"index": "global_mmlu_yo_medical__bfb0ac36__s000__tb262144__v002",
"uid": "global_mmlu_yo_medical__bfb0ac36__s000__tb262144__v002",
"axis": "domain",
"budget": 2048,
"token_budget": 262144,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_yo_medical__bfb0ac36__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 686.5803730750644, \"nll_base\": 6.531723294939313, \"eval_ref\": \"global_mmlu_yo_medical__bfb0ac36__s000__tb65536__v000\", \"oracle_delta\":... | {
"index": "global_mmlu_yo_medical__bfb0ac36__s000__tb65536__v000",
"uid": "global_mmlu_yo_medical__bfb0ac36__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_yo_medical__bfb0ac36__s001__tb65536__v003\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 752.6899664439626, \"nll_base\": 6.623653411865234, \"eval_ref\": \"global_mmlu_yo_medical__bfb0ac36__s001__tb65536__v003\", \"oracle_delta\":... | {
"index": "global_mmlu_yo_medical__bfb0ac36__s001__tb65536__v003",
"uid": "global_mmlu_yo_medical__bfb0ac36__s001__tb65536__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_zh_humanities__d05f5ebb__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 771.1803609898019, \"nll_base\": 6.6479222774505615, \"eval_ref\": \"global_mmlu_zh_humanities__d05f5ebb__s000__tb65536__v000\", \"oracle_d... | {
"index": "global_mmlu_zh_humanities__d05f5ebb__s000__tb65536__v000",
"uid": "global_mmlu_zh_humanities__d05f5ebb__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_zh_medical__563517b5__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 914.6952792064976, \"nll_base\": 6.818590981619699, \"eval_ref\": \"global_mmlu_zh_medical__563517b5__s000__tb131072__v001\", \"oracle_delta... | {
"index": "global_mmlu_zh_medical__563517b5__s000__tb131072__v001",
"uid": "global_mmlu_zh_medical__563517b5__s000__tb131072__v001",
"axis": "domain",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_zh_medical__563517b5__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 914.6952792064976, \"nll_base\": 6.818590981619699, \"eval_ref\": \"global_mmlu_zh_medical__563517b5__s000__tb65536__v000\", \"oracle_delta\":... | {
"index": "global_mmlu_zh_medical__563517b5__s000__tb65536__v000",
"uid": "global_mmlu_zh_medical__563517b5__s000__tb65536__v000",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_mmlu_zh_medical__563517b5__s001__tb65536__v003\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 925.2388931895225, \"nll_base\": 6.830051967075893, \"eval_ref\": \"global_mmlu_zh_medical__563517b5__s001__tb65536__v003\", \"oracle_delta\":... | {
"index": "global_mmlu_zh_medical__563517b5__s001__tb65536__v003",
"uid": "global_mmlu_zh_medical__563517b5__s001__tb65536__v003",
"axis": "domain",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_piqa_nonparallel_cloze_ben_latn__b41a431a__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2.8797833797341905, \"nll_base\": 1.057715075948964, \"eval_ref\": \"global_piqa_nonparallel_cloze_ben_latn__b41a431a__s000__t... | {
"index": "global_piqa_nonparallel_cloze_ben_latn__b41a431a__s000__tb65536__v000",
"uid": "global_piqa_nonparallel_cloze_ben_latn__b41a431a__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_piqa_nonparallel_cloze_btx_latn__dbb204ba__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2.782668196391163, \"nll_base\": 1.0234102502799816, \"eval_ref\": \"global_piqa_nonparallel_cloze_btx_latn__dbb204ba__s000__t... | {
"index": "global_piqa_nonparallel_cloze_btx_latn__dbb204ba__s000__tb65536__v000",
"uid": "global_piqa_nonparallel_cloze_btx_latn__dbb204ba__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_piqa_nonparallel_cloze_ckm_latn__8ba05f3b__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 12.241144906365866, \"nll_base\": 2.504802810811551, \"eval_ref\": \"global_piqa_nonparallel_cloze_ckm_latn__8ba05f3b__s000__t... | {
"index": "global_piqa_nonparallel_cloze_ckm_latn__8ba05f3b__s000__tb65536__v000",
"uid": "global_piqa_nonparallel_cloze_ckm_latn__8ba05f3b__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_piqa_nonparallel_generation_afb_arab__2f5d5b62__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2524.5466139089294, \"nll_base\": 7.833816766738892, \"eval_ref\": \"global_piqa_nonparallel_generation_afb_arab__2f5d5... | {
"index": "global_piqa_nonparallel_generation_afb_arab__2f5d5b62__s000__tb131072__v000",
"uid": "global_piqa_nonparallel_generation_afb_arab__2f5d5b62__s000__tb131072__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
datacuration | [
{
"role": "system",
"content": "You are a **training-data curator**. A small base language model needs to do better on a **target task**, and you must choose which examples from a large pool of candidate training examples it should be fine-tuned on.\n\nThe base model will be LoRA-fine-tuned on exactly the e... | datacuration_selection | {
"style": "datacuration",
"ground_truth": "{\"task_id\": \"global_piqa_nonparallel_generation_bho_deva__73eb00e2__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3666.107731638088, \"nll_base\": 8.206885814666748, \"eval_ref\": \"global_piqa_nonparallel_generation_bho_deva__73eb00e2... | {
"index": "global_piqa_nonparallel_generation_bho_deva__73eb00e2__s000__tb65536__v000",
"uid": "global_piqa_nonparallel_generation_bho_deva__73eb00e2__s000__tb65536__v000",
"axis": "multilingual",
"budget": 2048,
"token_budget": 65536,
"variant": "filter"
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.