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\": \"mmmlu_sw_ke_jurisprudence__33180790__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1073.8881213186262, \"nll_base\": 6.97904109954834, \"eval_ref\": \"mmmlu_sw_ke_jurisprudence__33180790__s000__tb131072__v000\", \"oracle... | {
"index": "mmmlu_sw_ke_jurisprudence__33180790__s000__tb131072__v000",
"uid": "mmmlu_sw_ke_jurisprudence__33180790__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\": \"mmmlu_sw_ke_medical_genetics__3b839168__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1417.6969518296596, \"nll_base\": 7.256788969039917, \"eval_ref\": \"mmmlu_sw_ke_medical_genetics__3b839168__s000__tb131072__v000\", \... | {
"index": "mmmlu_sw_ke_medical_genetics__3b839168__s000__tb131072__v000",
"uid": "mmmlu_sw_ke_medical_genetics__3b839168__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\": \"mmmlu_sw_ke_medical_genetics__3b839168__s000__tb262144__v001\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1417.6969518296596, \"nll_base\": 7.256788969039917, \"eval_ref\": \"mmmlu_sw_ke_medical_genetics__3b839168__s000__tb262144__v001\", \... | {
"index": "mmmlu_sw_ke_medical_genetics__3b839168__s000__tb262144__v001",
"uid": "mmmlu_sw_ke_medical_genetics__3b839168__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\": \"mmmlu_sw_ke_medical_genetics__3b839168__s000__tb65536__v002\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1417.6969518296596, \"nll_base\": 7.256788969039917, \"eval_ref\": \"mmmlu_sw_ke_medical_genetics__3b839168__s000__tb65536__v002\", \"or... | {
"index": "mmmlu_sw_ke_medical_genetics__3b839168__s000__tb65536__v002",
"uid": "mmmlu_sw_ke_medical_genetics__3b839168__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\": \"mmmlu_sw_ke_security_studies__a1357903__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1091.0275009753775, \"nll_base\": 6.994875192642212, \"eval_ref\": \"mmmlu_sw_ke_security_studies__a1357903__s000__tb262144__v000\", \... | {
"index": "mmmlu_sw_ke_security_studies__a1357903__s000__tb262144__v000",
"uid": "mmmlu_sw_ke_security_studies__a1357903__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\": \"mmmlu_yo_ng_anatomy__b10c741a__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 620.8275834835268, \"nll_base\": 6.431053400039673, \"eval_ref\": \"mmmlu_yo_ng_anatomy__b10c741a__s000__tb262144__v000\", \"oracle_delta\": 1.... | {
"index": "mmmlu_yo_ng_anatomy__b10c741a__s000__tb262144__v000",
"uid": "mmmlu_yo_ng_anatomy__b10c741a__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\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 846.3872937437898, \"nll_base\": 6.740977048873901, \"eval_ref\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb131072__v001\"... | {
"index": "mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb131072__v001",
"uid": "mmmlu_yo_ng_clinical_knowledge__01637c6c__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\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb262144__v002\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 846.3872937437898, \"nll_base\": 6.740977048873901, \"eval_ref\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb262144__v002\"... | {
"index": "mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb262144__v002",
"uid": "mmmlu_yo_ng_clinical_knowledge__01637c6c__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\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 846.3872937437898, \"nll_base\": 6.740977048873901, \"eval_ref\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb65536__v000\", \... | {
"index": "mmmlu_yo_ng_clinical_knowledge__01637c6c__s000__tb65536__v000",
"uid": "mmmlu_yo_ng_clinical_knowledge__01637c6c__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\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s001__tb65536__v003\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 929.3614061186418, \"nll_base\": 6.834497690200806, \"eval_ref\": \"mmmlu_yo_ng_clinical_knowledge__01637c6c__s001__tb65536__v003\", \... | {
"index": "mmmlu_yo_ng_clinical_knowledge__01637c6c__s001__tb65536__v003",
"uid": "mmmlu_yo_ng_clinical_knowledge__01637c6c__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\": \"mmmlu_yo_ng_high_school_biology__1ad0e994__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 567.7972779928407, \"nll_base\": 6.341764450073242, \"eval_ref\": \"mmmlu_yo_ng_high_school_biology__1ad0e994__s000__tb262144__v000... | {
"index": "mmmlu_yo_ng_high_school_biology__1ad0e994__s000__tb262144__v000",
"uid": "mmmlu_yo_ng_high_school_biology__1ad0e994__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\": \"mmmlu_yo_ng_high_school_macroeconomics__7863db0b__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 776.2837631021175, \"nll_base\": 6.654518127441406, \"eval_ref\": \"mmmlu_yo_ng_high_school_macroeconomics__7863db0b__s000__... | {
"index": "mmmlu_yo_ng_high_school_macroeconomics__7863db0b__s000__tb262144__v000",
"uid": "mmmlu_yo_ng_high_school_macroeconomics__7863db0b__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\": \"mmmlu_yo_ng_human_sexuality__59684003__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 858.8178252819113, \"nll_base\": 6.75555682182312, \"eval_ref\": \"mmmlu_yo_ng_human_sexuality__59684003__s000__tb131072__v000\", \"ora... | {
"index": "mmmlu_yo_ng_human_sexuality__59684003__s000__tb131072__v000",
"uid": "mmmlu_yo_ng_human_sexuality__59684003__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\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 724.159189936986, \"nll_base\": 6.58501124382019, \"eval_ref\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb131072__v000\", \"or... | {
"index": "mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb131072__v000",
"uid": "mmmlu_yo_ng_medical_genetics__0c4f7025__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\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb262144__v001\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 724.159189936986, \"nll_base\": 6.58501124382019, \"eval_ref\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb262144__v001\", \"or... | {
"index": "mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb262144__v001",
"uid": "mmmlu_yo_ng_medical_genetics__0c4f7025__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\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb65536__v002\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 724.159189936986, \"nll_base\": 6.58501124382019, \"eval_ref\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb65536__v002\", \"oracl... | {
"index": "mmmlu_yo_ng_medical_genetics__0c4f7025__s000__tb65536__v002",
"uid": "mmmlu_yo_ng_medical_genetics__0c4f7025__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\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s001__tb131072__v003\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 715.6325573879843, \"nll_base\": 6.573166847229004, \"eval_ref\": \"mmmlu_yo_ng_medical_genetics__0c4f7025__s001__tb131072__v003\", \"... | {
"index": "mmmlu_yo_ng_medical_genetics__0c4f7025__s001__tb131072__v003",
"uid": "mmmlu_yo_ng_medical_genetics__0c4f7025__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\": \"mmmlu_yo_ng_nutrition__d4a0ed61__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 820.3351548484336, \"nll_base\": 6.709712982177734, \"eval_ref\": \"mmmlu_yo_ng_nutrition__d4a0ed61__s000__tb131072__v000\", \"oracle_delta\"... | {
"index": "mmmlu_yo_ng_nutrition__d4a0ed61__s000__tb131072__v000",
"uid": "mmmlu_yo_ng_nutrition__d4a0ed61__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\": \"mmmlu_yo_ng_professional_law__ec1d79c3__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 484.43448792371277, \"nll_base\": 6.1829822063446045, \"eval_ref\": \"mmmlu_yo_ng_professional_law__ec1d79c3__s000__tb131072__v000\", ... | {
"index": "mmmlu_yo_ng_professional_law__ec1d79c3__s000__tb131072__v000",
"uid": "mmmlu_yo_ng_professional_law__ec1d79c3__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\": \"mmmlu_yo_ng_professional_psychology__acc76165__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 559.7117413490718, \"nll_base\": 6.3274219036102295, \"eval_ref\": \"mmmlu_yo_ng_professional_psychology__acc76165__s000__tb262... | {
"index": "mmmlu_yo_ng_professional_psychology__acc76165__s000__tb262144__v000",
"uid": "mmmlu_yo_ng_professional_psychology__acc76165__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\": \"mmmlu_yo_ng_virology__a269b563__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 577.4615516032911, \"nll_base\": 6.358641862869263, \"eval_ref\": \"mmmlu_yo_ng_virology__a269b563__s000__tb262144__v000\", \"oracle_delta\": ... | {
"index": "mmmlu_yo_ng_virology__a269b563__s000__tb262144__v000",
"uid": "mmmlu_yo_ng_virology__a269b563__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\": \"mmmlu_zh_cn_clinical_knowledge__7131726f__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1132.808746831704, \"nll_base\": 7.0324554443359375, \"eval_ref\": \"mmmlu_zh_cn_clinical_knowledge__7131726f__s000__tb131072__v000\... | {
"index": "mmmlu_zh_cn_clinical_knowledge__7131726f__s000__tb131072__v000",
"uid": "mmmlu_zh_cn_clinical_knowledge__7131726f__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\": \"mmmlu_zh_cn_clinical_knowledge__7131726f__s000__tb65536__v002\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1132.808746831704, \"nll_base\": 7.0324554443359375, \"eval_ref\": \"mmmlu_zh_cn_clinical_knowledge__7131726f__s000__tb65536__v002\", ... | {
"index": "mmmlu_zh_cn_clinical_knowledge__7131726f__s000__tb65536__v002",
"uid": "mmmlu_zh_cn_clinical_knowledge__7131726f__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\": \"mmmlu_zh_cn_clinical_knowledge__7131726f__s001__tb131072__v003\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1083.8833534543696, \"nll_base\": 6.988305568695068, \"eval_ref\": \"mmmlu_zh_cn_clinical_knowledge__7131726f__s001__tb131072__v003\... | {
"index": "mmmlu_zh_cn_clinical_knowledge__7131726f__s001__tb131072__v003",
"uid": "mmmlu_zh_cn_clinical_knowledge__7131726f__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\": \"mmmlu_zh_cn_college_mathematics__21667794__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 866.8007471473458, \"nll_base\": 6.7648091316223145, \"eval_ref\": \"mmmlu_zh_cn_college_mathematics__21667794__s000__tb131072__v00... | {
"index": "mmmlu_zh_cn_college_mathematics__21667794__s000__tb131072__v000",
"uid": "mmmlu_zh_cn_college_mathematics__21667794__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\": \"mmmlu_zh_cn_computer_security__3281c41d__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1062.129282753376, \"nll_base\": 6.96803092956543, \"eval_ref\": \"mmmlu_zh_cn_computer_security__3281c41d__s000__tb65536__v000\", \"or... | {
"index": "mmmlu_zh_cn_computer_security__3281c41d__s000__tb65536__v000",
"uid": "mmmlu_zh_cn_computer_security__3281c41d__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\": \"mmmlu_zh_cn_high_school_government_and_politics__a2ce0931__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1258.0220455356446, \"nll_base\": 7.137295961380005, \"eval_ref\": \"mmmlu_zh_cn_high_school_government_and_politics_... | {
"index": "mmmlu_zh_cn_high_school_government_and_politics__a2ce0931__s000__tb65536__v000",
"uid": "mmmlu_zh_cn_high_school_government_and_politics__a2ce0931__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\": \"mmmlu_zh_cn_high_school_physics__b57bef62__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1236.7123904701934, \"nll_base\": 7.120211839675903, \"eval_ref\": \"mmmlu_zh_cn_high_school_physics__b57bef62__s000__tb65536__v000\"... | {
"index": "mmmlu_zh_cn_high_school_physics__b57bef62__s000__tb65536__v000",
"uid": "mmmlu_zh_cn_high_school_physics__b57bef62__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\": \"mmmlu_zh_cn_marketing__645adacc__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 880.1261927824228, \"nll_base\": 6.780065298080444, \"eval_ref\": \"mmmlu_zh_cn_marketing__645adacc__s000__tb131072__v000\", \"oracle_delta\"... | {
"index": "mmmlu_zh_cn_marketing__645adacc__s000__tb131072__v000",
"uid": "mmmlu_zh_cn_marketing__645adacc__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\": \"mmmlu_zh_cn_medical_genetics__ecbc982d__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 812.8260919145355, \"nll_base\": 6.700517177581787, \"eval_ref\": \"mmmlu_zh_cn_medical_genetics__ecbc982d__s000__tb131072__v000\", \"... | {
"index": "mmmlu_zh_cn_medical_genetics__ecbc982d__s000__tb131072__v000",
"uid": "mmmlu_zh_cn_medical_genetics__ecbc982d__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\": \"mmmlu_zh_cn_medical_genetics__ecbc982d__s000__tb65536__v002\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 812.8260919145355, \"nll_base\": 6.700517177581787, \"eval_ref\": \"mmmlu_zh_cn_medical_genetics__ecbc982d__s000__tb65536__v002\", \"ora... | {
"index": "mmmlu_zh_cn_medical_genetics__ecbc982d__s000__tb65536__v002",
"uid": "mmmlu_zh_cn_medical_genetics__ecbc982d__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\": \"mnli__8a11be60__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 929.7561185979673, \"nll_base\": 6.8349223136901855, \"eval_ref\": \"mnli__8a11be60__s000__tb131072__v000\", \"oracle_delta\": 0.42839622497558594, \"random_d... | {
"index": "mnli__8a11be60__s000__tb131072__v000",
"uid": "mnli__8a11be60__s000__tb131072__v000",
"axis": "instruction",
"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\": \"mnli_mismatch__01f44233__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 777.9062058977089, \"nll_base\": 6.656605958938599, \"eval_ref\": \"mnli_mismatch__01f44233__s000__tb65536__v000\", \"oracle_delta\": 0.487712144851684... | {
"index": "mnli_mismatch__01f44233__s000__tb65536__v000",
"uid": "mnli_mismatch__01f44233__s000__tb65536__v000",
"axis": "instruction",
"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\": \"multiblimp_azz__9e64d3d9__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1.5499730407025691, \"nll_base\": 0.43823753768477697, \"eval_ref\": \"multiblimp_azz__9e64d3d9__s000__tb65536__v000\", \"oracle_delta\": 0.0228778000... | {
"index": "multiblimp_azz__9e64d3d9__s000__tb65536__v000",
"uid": "multiblimp_azz__9e64d3d9__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\": \"multiblimp_fra__1b353488__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1.3659454286176336, \"nll_base\": 0.3118468105831134, \"eval_ref\": \"multiblimp_fra__1b353488__s000__tb262144__v000\", \"oracle_delta\": 0.05725127... | {
"index": "multiblimp_fra__1b353488__s000__tb262144__v000",
"uid": "multiblimp_fra__1b353488__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\": \"multiblimp_fro__f7b5785e__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2.102723619354202, \"nll_base\": 0.7432334659984277, \"eval_ref\": \"multiblimp_fro__f7b5785e__s000__tb131072__v000\", \"oracle_delta\": 0.109810576... | {
"index": "multiblimp_fro__f7b5785e__s000__tb131072__v000",
"uid": "multiblimp_fro__f7b5785e__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\": \"multiblimp_kmr__40389893__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1.651465211411837, \"nll_base\": 0.5016629007838717, \"eval_ref\": \"multiblimp_kmr__40389893__s000__tb65536__v000\", \"oracle_delta\": 0.011129260406... | {
"index": "multiblimp_kmr__40389893__s000__tb65536__v000",
"uid": "multiblimp_kmr__40389893__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\": \"multiblimp_por__3d2f138c__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1.5094754617387989, \"nll_base\": 0.4117622141457774, \"eval_ref\": \"multiblimp_por__3d2f138c__s000__tb131072__v000\", \"oracle_delta\": 0.06799903... | {
"index": "multiblimp_por__3d2f138c__s000__tb131072__v000",
"uid": "multiblimp_por__3d2f138c__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\": \"naijarc_hau_prompt_5__8de7bfdc__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 941.5484762019679, \"nll_base\": 6.8475258350372314, \"eval_ref\": \"naijarc_hau_prompt_5__8de7bfdc__s000__tb65536__v000\", \"oracle_delta\": 0.... | {
"index": "naijarc_hau_prompt_5__8de7bfdc__s000__tb65536__v000",
"uid": "naijarc_hau_prompt_5__8de7bfdc__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\": \"naijarc_yor_prompt_1__fe34adbf__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1558.057816828524, \"nll_base\": 7.351195335388184, \"eval_ref\": \"naijarc_yor_prompt_1__fe34adbf__s000__tb262144__v000\", \"oracle_delta\": ... | {
"index": "naijarc_yor_prompt_1__fe34adbf__s000__tb262144__v000",
"uid": "naijarc_yor_prompt_1__fe34adbf__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\": \"naijarc_yor_prompt_4__76ab309f__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 539.3316287578792, \"nll_base\": 6.290330648422241, \"eval_ref\": \"naijarc_yor_prompt_4__76ab309f__s000__tb65536__v000\", \"oracle_delta\": 0.4... | {
"index": "naijarc_yor_prompt_4__76ab309f__s000__tb65536__v000",
"uid": "naijarc_yor_prompt_4__76ab309f__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\": \"non_greedy_robustness_agieval_aqua_rat__f7e594ba__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2239.3931002395243, \"nll_base\": 7.71396017074585, \"eval_ref\": \"non_greedy_robustness_agieval_aqua_rat__f7e594ba__s000__tb... | {
"index": "non_greedy_robustness_agieval_aqua_rat__f7e594ba__s000__tb65536__v000",
"uid": "non_greedy_robustness_agieval_aqua_rat__f7e594ba__s000__tb65536__v000",
"axis": "instruction",
"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\": \"non_greedy_robustness_agieval_logiqa_en__ea4afef2__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2024.700969997641, \"nll_base\": 7.613177299499512, \"eval_ref\": \"non_greedy_robustness_agieval_logiqa_en__ea4afef2__s000__... | {
"index": "non_greedy_robustness_agieval_logiqa_en__ea4afef2__s000__tb65536__v000",
"uid": "non_greedy_robustness_agieval_logiqa_en__ea4afef2__s000__tb65536__v000",
"axis": "instruction",
"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\": \"non_greedy_robustness_agieval_lsat_lr__5836698e__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2689.0380349448587, \"nll_base\": 7.896938800811768, \"eval_ref\": \"non_greedy_robustness_agieval_lsat_lr__5836698e__s000__tb6... | {
"index": "non_greedy_robustness_agieval_lsat_lr__5836698e__s000__tb65536__v000",
"uid": "non_greedy_robustness_agieval_lsat_lr__5836698e__s000__tb65536__v000",
"axis": "instruction",
"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\": \"non_greedy_robustness_agieval_sat_en__66cb9fa8__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1028.5324818925687, \"nll_base\": 6.935888290405273, \"eval_ref\": \"non_greedy_robustness_agieval_sat_en__66cb9fa8__s000__tb2... | {
"index": "non_greedy_robustness_agieval_sat_en__66cb9fa8__s000__tb262144__v000",
"uid": "non_greedy_robustness_agieval_sat_en__66cb9fa8__s000__tb262144__v000",
"axis": "instruction",
"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\": \"non_greedy_robustness_math_counting_and_prob__76d66df3__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 72.14328186577292, \"nll_base\": 4.278654167451054, \"eval_ref\": \"non_greedy_robustness_math_counting_and_prob__76d6... | {
"index": "non_greedy_robustness_math_counting_and_prob__76d66df3__s000__tb262144__v000",
"uid": "non_greedy_robustness_math_counting_and_prob__76d66df3__s000__tb262144__v000",
"axis": "instruction",
"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\": \"non_greedy_robustness_math_geometry__923480e4__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 73.73853560715547, \"nll_base\": 4.300525533741918, \"eval_ref\": \"non_greedy_robustness_math_geometry__923480e4__s000__tb1310... | {
"index": "non_greedy_robustness_math_geometry__923480e4__s000__tb131072__v000",
"uid": "non_greedy_robustness_math_geometry__923480e4__s000__tb131072__v000",
"axis": "instruction",
"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\": \"non_greedy_robustness_math_intermediate_algebra__814ef752__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 68.07495239600512, \"nll_base\": 4.220609339272103, \"eval_ref\": \"non_greedy_robustness_math_intermediate_algebra... | {
"index": "non_greedy_robustness_math_intermediate_algebra__814ef752__s000__tb131072__v000",
"uid": "non_greedy_robustness_math_intermediate_algebra__814ef752__s000__tb131072__v000",
"axis": "instruction",
"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\": \"non_greedy_robustness_math_prealgebra__e429225e__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 99.53829107631172, \"nll_base\": 4.600542405072381, \"eval_ref\": \"non_greedy_robustness_math_prealgebra__e429225e__s000__tb... | {
"index": "non_greedy_robustness_math_prealgebra__e429225e__s000__tb131072__v000",
"uid": "non_greedy_robustness_math_prealgebra__e429225e__s000__tb131072__v000",
"axis": "instruction",
"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\": \"noor_abstract_algebra__2b025ae0__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 48.29608390865984, \"nll_base\": 3.8773504788758326, \"eval_ref\": \"noor_abstract_algebra__2b025ae0__s000__tb131072__v000\", \"oracle_delta\... | {
"index": "noor_abstract_algebra__2b025ae0__s000__tb131072__v000",
"uid": "noor_abstract_algebra__2b025ae0__s000__tb131072__v000",
"axis": "instruction",
"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\": \"noor_college_mathematics__d52eedc5__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 15.31667844868018, \"nll_base\": 2.728942329361594, \"eval_ref\": \"noor_college_mathematics__d52eedc5__s000__tb65536__v000\", \"oracle_delt... | {
"index": "noor_college_mathematics__d52eedc5__s000__tb65536__v000",
"uid": "noor_college_mathematics__d52eedc5__s000__tb65536__v000",
"axis": "instruction",
"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\": \"noor_elementary_mathematics__846cbd70__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 12.20011461550274, \"nll_base\": 2.50144534640842, \"eval_ref\": \"noor_elementary_mathematics__846cbd70__s000__tb65536__v000\", \"oracle... | {
"index": "noor_elementary_mathematics__846cbd70__s000__tb65536__v000",
"uid": "noor_elementary_mathematics__846cbd70__s000__tb65536__v000",
"axis": "instruction",
"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\": \"noor_high_school_statistics__1d6a04bb__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3.0968559952547814, \"nll_base\": 1.130387401768173, \"eval_ref\": \"noor_high_school_statistics__1d6a04bb__s000__tb131072__v000\", \"o... | {
"index": "noor_high_school_statistics__1d6a04bb__s000__tb131072__v000",
"uid": "noor_high_school_statistics__1d6a04bb__s000__tb131072__v000",
"axis": "instruction",
"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\": \"norcommonsenseqa_nno_p4__57ac618a__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1407.5973335309047, \"nll_base\": 7.249639511108398, \"eval_ref\": \"norcommonsenseqa_nno_p4__57ac618a__s000__tb262144__v000\", \"oracle_de... | {
"index": "norcommonsenseqa_nno_p4__57ac618a__s000__tb262144__v000",
"uid": "norcommonsenseqa_nno_p4__57ac618a__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\": \"noridiom_nno_p1__1f7930b5__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2126.096473238812, \"nll_base\": 7.66204293568929, \"eval_ref\": \"noridiom_nno_p1__1f7930b5__s000__tb131072__v000\", \"oracle_delta\": 0.194834709... | {
"index": "noridiom_nno_p1__1f7930b5__s000__tb131072__v000",
"uid": "noridiom_nno_p1__1f7930b5__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\": \"noropenbookqa_nno_p1__97576374__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 5.098487346653628, \"nll_base\": 1.6289438970408208, \"eval_ref\": \"noropenbookqa_nno_p1__97576374__s000__tb131072__v000\", \"oracle_delta\":... | {
"index": "noropenbookqa_nno_p1__97576374__s000__tb131072__v000",
"uid": "noropenbookqa_nno_p1__97576374__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\": \"noropenbookqa_nno_p2__cba6ec95__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1655.9850989231732, \"nll_base\": 7.412151336669922, \"eval_ref\": \"noropenbookqa_nno_p2__cba6ec95__s000__tb262144__v000\", \"oracle_delta\":... | {
"index": "noropenbookqa_nno_p2__cba6ec95__s000__tb262144__v000",
"uid": "noropenbookqa_nno_p2__cba6ec95__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\": \"noropenbookqa_nno_p3__61e77641__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3339.4332469828714, \"nll_base\": 8.113556385040283, \"eval_ref\": \"noropenbookqa_nno_p3__61e77641__s000__tb262144__v000\", \"oracle_delta\":... | {
"index": "noropenbookqa_nno_p3__61e77641__s000__tb262144__v000",
"uid": "noropenbookqa_nno_p3__61e77641__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\": \"norquad_p4__c9b5a673__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 15.677854034443309, \"nll_base\": 2.7522491455078124, \"eval_ref\": \"norquad_p4__c9b5a673__s000__tb65536__v000\", \"oracle_delta\": 0.08425968715122778, ... | {
"index": "norquad_p4__c9b5a673__s000__tb65536__v000",
"uid": "norquad_p4__c9b5a673__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\": \"noticia__3498034b__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 7.851561337043888, \"nll_base\": 2.0607124084557547, \"eval_ref\": \"noticia__3498034b__s000__tb262144__v000\", \"oracle_delta\": 0.09436782179024705, \"ra... | {
"index": "noticia__3498034b__s000__tb262144__v000",
"uid": "noticia__3498034b__s000__tb262144__v000",
"axis": "instruction",
"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\": \"nq_open__48f3b8e7__s000__tb131072__v002\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 199.41338695868728, \"nll_base\": 5.2953799914762465, \"eval_ref\": \"nq_open__48f3b8e7__s000__tb131072__v002\", \"oracle_delta\": 0.6146974067467488, \"ra... | {
"index": "nq_open__48f3b8e7__s000__tb131072__v002",
"uid": "nq_open__48f3b8e7__s000__tb131072__v002",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 199.41338695868728, \"nll_base\": 5.2953799914762465, \"eval_ref\": \"nq_open__48f3b8e7__s000__tb262144__v000\", \"oracle_delta\": 0.806855659264361, \"ran... | {
"index": "nq_open__48f3b8e7__s000__tb262144__v000",
"uid": "nq_open__48f3b8e7__s000__tb262144__v000",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s000__tb65536__v001\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 199.41338695868728, \"nll_base\": 5.2953799914762465, \"eval_ref\": \"nq_open__48f3b8e7__s000__tb65536__v001\", \"oracle_delta\": 0.28898576504922335, \"rand... | {
"index": "nq_open__48f3b8e7__s000__tb65536__v001",
"uid": "nq_open__48f3b8e7__s000__tb65536__v001",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s001__tb131072__v005\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 193.8366375106093, \"nll_base\": 5.267015729631696, \"eval_ref\": \"nq_open__48f3b8e7__s001__tb131072__v005\", \"oracle_delta\": 0.4651078723725819, \"rand... | {
"index": "nq_open__48f3b8e7__s001__tb131072__v005",
"uid": "nq_open__48f3b8e7__s001__tb131072__v005",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s001__tb262144__v003\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 193.8366375106093, \"nll_base\": 5.267015729631696, \"eval_ref\": \"nq_open__48f3b8e7__s001__tb262144__v003\", \"oracle_delta\": 0.6390248253231956, \"rand... | {
"index": "nq_open__48f3b8e7__s001__tb262144__v003",
"uid": "nq_open__48f3b8e7__s001__tb262144__v003",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s001__tb65536__v004\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 193.8366375106093, \"nll_base\": 5.267015729631696, \"eval_ref\": \"nq_open__48f3b8e7__s001__tb65536__v004\", \"oracle_delta\": 0.3239250183105469, \"random_... | {
"index": "nq_open__48f3b8e7__s001__tb65536__v004",
"uid": "nq_open__48f3b8e7__s001__tb65536__v004",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s002__tb131072__v008\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 177.46921528330512, \"nll_base\": 5.178797158850244, \"eval_ref\": \"nq_open__48f3b8e7__s002__tb131072__v008\", \"oracle_delta\": 0.6129791995128953, \"ran... | {
"index": "nq_open__48f3b8e7__s002__tb131072__v008",
"uid": "nq_open__48f3b8e7__s002__tb131072__v008",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s002__tb262144__v006\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 177.46921528330512, \"nll_base\": 5.178797158850244, \"eval_ref\": \"nq_open__48f3b8e7__s002__tb262144__v006\", \"oracle_delta\": 0.7429248855774659, \"ran... | {
"index": "nq_open__48f3b8e7__s002__tb262144__v006",
"uid": "nq_open__48f3b8e7__s002__tb262144__v006",
"axis": "qa",
"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\": \"nq_open__48f3b8e7__s002__tb65536__v007\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 177.46921528330512, \"nll_base\": 5.178797158850244, \"eval_ref\": \"nq_open__48f3b8e7__s002__tb65536__v007\", \"oracle_delta\": 0.31076638095350084, \"rando... | {
"index": "nq_open__48f3b8e7__s002__tb65536__v007",
"uid": "nq_open__48f3b8e7__s002__tb65536__v007",
"axis": "qa",
"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\": \"nrk_quiz_qa_nob_p4__db0e86cd__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 6.645614878163226, \"nll_base\": 1.8939572203393076, \"eval_ref\": \"nrk_quiz_qa_nob_p4__db0e86cd__s000__tb65536__v000\", \"oracle_delta\": 0.0899... | {
"index": "nrk_quiz_qa_nob_p4__db0e86cd__s000__tb65536__v000",
"uid": "nrk_quiz_qa_nob_p4__db0e86cd__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\": \"ntrex_afr_Latn-eng_Latn_prompt_3__2243819d__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 34.5208676407375, \"nll_base\": 3.541564000335559, \"eval_ref\": \"ntrex_afr_Latn-eng_Latn_prompt_3__2243819d__s000__tb65536__v000\"... | {
"index": "ntrex_afr_Latn-eng_Latn_prompt_3__2243819d__s000__tb65536__v000",
"uid": "ntrex_afr_Latn-eng_Latn_prompt_3__2243819d__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\": \"ntrex_bem_Latn-eng_Latn_prompt_1__ffc6c799__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 34.40805028859724, \"nll_base\": 3.538290557018301, \"eval_ref\": \"ntrex_bem_Latn-eng_Latn_prompt_1__ffc6c799__s000__tb65536__v000\... | {
"index": "ntrex_bem_Latn-eng_Latn_prompt_1__ffc6c799__s000__tb65536__v000",
"uid": "ntrex_bem_Latn-eng_Latn_prompt_1__ffc6c799__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\": \"ntrex_eng_Latn-fra_Latn_prompt_1__fd880140__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 19.52347282109358, \"nll_base\": 2.9716174760773555, \"eval_ref\": \"ntrex_eng_Latn-fra_Latn_prompt_1__fd880140__s000__tb131072__v... | {
"index": "ntrex_eng_Latn-fra_Latn_prompt_1__fd880140__s000__tb131072__v000",
"uid": "ntrex_eng_Latn-fra_Latn_prompt_1__fd880140__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\": \"ntrex_eng_Latn-kin_Latn_prompt_2__3d0b2c9d__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 297.8289746652329, \"nll_base\": 5.696519411236372, \"eval_ref\": \"ntrex_eng_Latn-kin_Latn_prompt_2__3d0b2c9d__s000__tb65536__v000\... | {
"index": "ntrex_eng_Latn-kin_Latn_prompt_2__3d0b2c9d__s000__tb65536__v000",
"uid": "ntrex_eng_Latn-kin_Latn_prompt_2__3d0b2c9d__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\": \"ntrex_eng_Latn-mlg_Latn_prompt_1__0f466af0__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 445.1243844073812, \"nll_base\": 6.0983537586312595, \"eval_ref\": \"ntrex_eng_Latn-mlg_Latn_prompt_1__0f466af0__s000__tb262144__v... | {
"index": "ntrex_eng_Latn-mlg_Latn_prompt_1__0f466af0__s000__tb262144__v000",
"uid": "ntrex_eng_Latn-mlg_Latn_prompt_1__0f466af0__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\": \"ntrex_eng_Latn-tir_Ethi_prompt_2__8736d768__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 6.316759168578045, \"nll_base\": 1.843206286855167, \"eval_ref\": \"ntrex_eng_Latn-tir_Ethi_prompt_2__8736d768__s000__tb131072__v0... | {
"index": "ntrex_eng_Latn-tir_Ethi_prompt_2__8736d768__s000__tb131072__v000",
"uid": "ntrex_eng_Latn-tir_Ethi_prompt_2__8736d768__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\": \"ntrex_eng_Latn-ton_Latn_prompt_2__41f6c230__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 92.50329026053522, \"nll_base\": 4.527244214269933, \"eval_ref\": \"ntrex_eng_Latn-ton_Latn_prompt_2__41f6c230__s000__tb65536__v000\... | {
"index": "ntrex_eng_Latn-ton_Latn_prompt_2__41f6c230__s000__tb65536__v000",
"uid": "ntrex_eng_Latn-ton_Latn_prompt_2__41f6c230__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\": \"ntrex_eng_Latn-ven_Latn_prompt_3__9a9a3c5d__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 223.1206267686715, \"nll_base\": 5.4077125524255045, \"eval_ref\": \"ntrex_eng_Latn-ven_Latn_prompt_3__9a9a3c5d__s000__tb131072__v... | {
"index": "ntrex_eng_Latn-ven_Latn_prompt_3__9a9a3c5d__s000__tb131072__v000",
"uid": "ntrex_eng_Latn-ven_Latn_prompt_3__9a9a3c5d__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\": \"ntrex_eng_Latn-wol_Latn_prompt_3__633212c2__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 595.0993643082439, \"nll_base\": 6.388728390440261, \"eval_ref\": \"ntrex_eng_Latn-wol_Latn_prompt_3__633212c2__s000__tb65536__v000\... | {
"index": "ntrex_eng_Latn-wol_Latn_prompt_3__633212c2__s000__tb65536__v000",
"uid": "ntrex_eng_Latn-wol_Latn_prompt_3__633212c2__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\": \"ntrex_sna_Latn-eng_Latn_prompt_2__d2d56f9b__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 46.89341435197968, \"nll_base\": 3.847877246658237, \"eval_ref\": \"ntrex_sna_Latn-eng_Latn_prompt_2__d2d56f9b__s000__tb65536__v000\... | {
"index": "ntrex_sna_Latn-eng_Latn_prompt_2__d2d56f9b__s000__tb65536__v000",
"uid": "ntrex_sna_Latn-eng_Latn_prompt_2__d2d56f9b__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\": \"nug__5261670b__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1.4691595042418983, \"nll_base\": 0.38469047144839635, \"eval_ref\": \"nug__5261670b__s000__tb131072__v001\", \"oracle_delta\": 0.029617004394531243, \"random_... | {
"index": "nug__5261670b__s000__tb131072__v001",
"uid": "nug__5261670b__s000__tb131072__v001",
"axis": "qa",
"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\": \"nug__5261670b__s000__tb262144__v002\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1.4691595042418983, \"nll_base\": 0.38469047144839635, \"eval_ref\": \"nug__5261670b__s000__tb262144__v002\", \"oracle_delta\": 0.04425562808388156, \"random_d... | {
"index": "nug__5261670b__s000__tb262144__v002",
"uid": "nug__5261670b__s000__tb262144__v002",
"axis": "qa",
"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\": \"nug__5261670b__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1.4691595042418983, \"nll_base\": 0.38469047144839635, \"eval_ref\": \"nug__5261670b__s000__tb65536__v000\", \"oracle_delta\": 0.015573104055304265, \"random_del... | {
"index": "nug__5261670b__s000__tb65536__v000",
"uid": "nug__5261670b__s000__tb65536__v000",
"axis": "qa",
"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\": \"nug__5261670b__s001__tb131072__v004\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1.496727023646013, \"nll_base\": 0.403280739876044, \"eval_ref\": \"nug__5261670b__s001__tb131072__v004\", \"oracle_delta\": 0.028283867668592477, \"random_del... | {
"index": "nug__5261670b__s001__tb131072__v004",
"uid": "nug__5261670b__s001__tb131072__v004",
"axis": "qa",
"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\": \"nug__5261670b__s001__tb262144__v005\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1.496727023646013, \"nll_base\": 0.403280739876044, \"eval_ref\": \"nug__5261670b__s001__tb262144__v005\", \"oracle_delta\": 0.04437184144794465, \"random_delt... | {
"index": "nug__5261670b__s001__tb262144__v005",
"uid": "nug__5261670b__s001__tb262144__v005",
"axis": "qa",
"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\": \"nug__5261670b__s002__tb131072__v007\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1.608813682054556, \"nll_base\": 0.47549706394943847, \"eval_ref\": \"nug__5261670b__s002__tb131072__v007\", \"oracle_delta\": 0.03229661897049685, \"random_de... | {
"index": "nug__5261670b__s002__tb131072__v007",
"uid": "nug__5261670b__s002__tb131072__v007",
"axis": "qa",
"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\": \"nug__5261670b__s002__tb262144__v008\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1.608813682054556, \"nll_base\": 0.47549706394943847, \"eval_ref\": \"nug__5261670b__s002__tb262144__v008\", \"oracle_delta\": 0.05093353438933368, \"random_de... | {
"index": "nug__5261670b__s002__tb262144__v008",
"uid": "nug__5261670b__s002__tb262144__v008",
"axis": "qa",
"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\": \"openai_mmlu_swa_prompt_2__c6b1182e__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 1665.4541316880454, \"nll_base\": 7.417853116989136, \"eval_ref\": \"openai_mmlu_swa_prompt_2__c6b1182e__s000__tb131072__v000\", \"oracle_... | {
"index": "openai_mmlu_swa_prompt_2__c6b1182e__s000__tb131072__v000",
"uid": "openai_mmlu_swa_prompt_2__c6b1182e__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\": \"openai_mmlu_swa_prompt_5__840d07c9__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 664.964520397709, \"nll_base\": 6.4997336864471436, \"eval_ref\": \"openai_mmlu_swa_prompt_5__840d07c9__s000__tb262144__v000\", \"oracle_d... | {
"index": "openai_mmlu_swa_prompt_5__840d07c9__s000__tb262144__v000",
"uid": "openai_mmlu_swa_prompt_5__840d07c9__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\": \"openai_mmlu_yor_prompt_3__5f6994b0__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4752.177365874373, \"nll_base\": 8.466358184814453, \"eval_ref\": \"openai_mmlu_yor_prompt_3__5f6994b0__s000__tb262144__v000\", \"oracle_d... | {
"index": "openai_mmlu_yor_prompt_3__5f6994b0__s000__tb262144__v000",
"uid": "openai_mmlu_yor_prompt_3__5f6994b0__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\": \"openbookqa__51283af1__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 21.952592360465914, \"nll_base\": 3.088885235530074, \"eval_ref\": \"openbookqa__51283af1__s000__tb131072__v000\", \"oracle_delta\": 0.15751861244119647... | {
"index": "openbookqa__51283af1__s000__tb131072__v000",
"uid": "openbookqa__51283af1__s000__tb131072__v000",
"axis": "knowledge",
"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\": \"persona_anti-LGBTQ-rights__1333a24f__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 8091.95595554531, \"nll_base\": 8.998625755310059, \"eval_ref\": \"persona_anti-LGBTQ-rights__1333a24f__s000__tb131072__v000\", \"oracle_... | {
"index": "persona_anti-LGBTQ-rights__1333a24f__s000__tb131072__v000",
"uid": "persona_anti-LGBTQ-rights__1333a24f__s000__tb131072__v000",
"axis": "instruction",
"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\": \"persona_anti-immigration__69c84931__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 6936.929398226827, \"nll_base\": 8.844614505767822, \"eval_ref\": \"persona_anti-immigration__69c84931__s000__tb65536__v000\", \"oracle_delt... | {
"index": "persona_anti-immigration__69c84931__s000__tb65536__v000",
"uid": "persona_anti-immigration__69c84931__s000__tb65536__v000",
"axis": "instruction",
"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\": \"persona_believes-it-is-not-being-watched-by-humans__78326f3a__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 8104.77647127963, \"nll_base\": 9.000208854675293, \"eval_ref\": \"persona_believes-it-is-not-being-watched-by-hum... | {
"index": "persona_believes-it-is-not-being-watched-by-humans__78326f3a__s000__tb65536__v000",
"uid": "persona_believes-it-is-not-being-watched-by-humans__78326f3a__s000__tb65536__v000",
"axis": "instruction",
"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\": \"persona_believes-it-knows-better-than-humans-about-how-the-system-should-behave__f25dfdad__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 9361.631383675778, \"nll_base\": 9.14437484741211, \"eval_ref\": \"persona_believes... | {
"index": "persona_believes-it-knows-better-than-humans-about-how-the-system-should-behave__f25dfdad__s000__tb131072__v000",
"uid": "persona_believes-it-knows-better-than-humans-about-how-the-system-should-behave__f25dfdad__s000__tb131072__v000",
"axis": "instruction",
"budget": 2048,
"token_budget": 131072,... |
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\": \"persona_believes-it-knows-better-than-humans-about-how-to-help-humans__454960ea__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 8962.857595933081, \"nll_base\": 9.100844383239746, \"eval_ref\": \"persona_believes-it-knows-b... | {
"index": "persona_believes-it-knows-better-than-humans-about-how-to-help-humans__454960ea__s000__tb65536__v000",
"uid": "persona_believes-it-knows-better-than-humans-about-how-to-help-humans__454960ea__s000__tb65536__v000",
"axis": "instruction",
"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\": \"persona_believes-life-has-no-meaning__aadac1ab__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 9583.658923419029, \"nll_base\": 9.1678147315979, \"eval_ref\": \"persona_believes-life-has-no-meaning__aadac1ab__s000__tb65536_... | {
"index": "persona_believes-life-has-no-meaning__aadac1ab__s000__tb65536__v000",
"uid": "persona_believes-life-has-no-meaning__aadac1ab__s000__tb65536__v000",
"axis": "instruction",
"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\": \"persona_cognitive-enhancement__065acb51__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 8831.276100869305, \"nll_base\": 9.086054801940918, \"eval_ref\": \"persona_cognitive-enhancement__065acb51__s000__tb131072__v000\", ... | {
"index": "persona_cognitive-enhancement__065acb51__s000__tb131072__v000",
"uid": "persona_cognitive-enhancement__065acb51__s000__tb131072__v000",
"axis": "instruction",
"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\": \"persona_desire-for-acquiring-compute__7647fb6d__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 7290.101909499848, \"nll_base\": 8.894272804260254, \"eval_ref\": \"persona_desire-for-acquiring-compute__7647fb6d__s000__tb6553... | {
"index": "persona_desire-for-acquiring-compute__7647fb6d__s000__tb65536__v000",
"uid": "persona_desire-for-acquiring-compute__7647fb6d__s000__tb65536__v000",
"axis": "instruction",
"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\": \"persona_desire-for-computational-efficiency__4f4c893d__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 7191.628582488499, \"nll_base\": 8.880672931671143, \"eval_ref\": \"persona_desire-for-computational-efficiency__4f4c893d... | {
"index": "persona_desire-for-computational-efficiency__4f4c893d__s000__tb65536__v000",
"uid": "persona_desire-for-computational-efficiency__4f4c893d__s000__tb65536__v000",
"axis": "instruction",
"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.