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_piqa_nonparallel_generation_ces_latn__20dfa4f9__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4467.01504301505, \"nll_base\": 8.404475688934326, \"eval_ref\": \"global_piqa_nonparallel_generation_ces_latn__20dfa4f... | {
"index": "global_piqa_nonparallel_generation_ces_latn__20dfa4f9__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_ces_latn__20dfa4f9__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_piqa_nonparallel_generation_dje_latn__40cf9756__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2888.4735206850955, \"nll_base\": 7.9684834480285645, \"eval_ref\": \"global_piqa_nonparallel_generation_dje_latn__40cf... | {
"index": "global_piqa_nonparallel_generation_dje_latn__40cf9756__s000__tb131072__v000",
"uid": "global_piqa_nonparallel_generation_dje_latn__40cf9756__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_fra_latn_fran__641b4e38__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3415.213840424033, \"nll_base\": 8.135995388031006, \"eval_ref\": \"global_piqa_nonparallel_generation_fra_latn_fr... | {
"index": "global_piqa_nonparallel_generation_fra_latn_fran__641b4e38__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_fra_latn_fran__641b4e38__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_piqa_nonparallel_generation_guj_gujr__0db84efb__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 5124.775426336823, \"nll_base\": 8.541841983795166, \"eval_ref\": \"global_piqa_nonparallel_generation_guj_gujr__0db84e... | {
"index": "global_piqa_nonparallel_generation_guj_gujr__0db84efb__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_guj_gujr__0db84efb__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_piqa_nonparallel_generation_lin_latn__393dd09f__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 2623.7384292506586, \"nll_base\": 7.8723554611206055, \"eval_ref\": \"global_piqa_nonparallel_generation_lin_latn__393d... | {
"index": "global_piqa_nonparallel_generation_lin_latn__393dd09f__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_lin_latn__393dd09f__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_piqa_nonparallel_generation_nob_latn__ac6c642b__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3810.6640287979894, \"nll_base\": 8.245558738708496, \"eval_ref\": \"global_piqa_nonparallel_generation_nob_latn__ac6c6... | {
"index": "global_piqa_nonparallel_generation_nob_latn__ac6c642b__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_nob_latn__ac6c642b__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_piqa_nonparallel_generation_slv_latn_prle__6d808c9b__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3419.8435465842867, \"nll_base\": 8.137350082397461, \"eval_ref\": \"global_piqa_nonparallel_generation_slv_latn_p... | {
"index": "global_piqa_nonparallel_generation_slv_latn_prle__6d808c9b__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_slv_latn_prle__6d808c9b__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_piqa_nonparallel_generation_spa_latn_mexi__75f460b1__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4266.042364363681, \"nll_base\": 8.358441829681396, \"eval_ref\": \"global_piqa_nonparallel_generation_spa_latn_me... | {
"index": "global_piqa_nonparallel_generation_spa_latn_mexi__75f460b1__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_spa_latn_mexi__75f460b1__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_piqa_nonparallel_generation_srp_cyrl__7e52728d__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4462.805940675945, \"nll_base\": 8.403532981872559, \"eval_ref\": \"global_piqa_nonparallel_generation_srp_cyrl__7e52728d... | {
"index": "global_piqa_nonparallel_generation_srp_cyrl__7e52728d__s000__tb65536__v000",
"uid": "global_piqa_nonparallel_generation_srp_cyrl__7e52728d__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_srp_latn__dcdf0a40__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3247.361150905348, \"nll_base\": 8.08559799194336, \"eval_ref\": \"global_piqa_nonparallel_generation_srp_latn__dcdf0a4... | {
"index": "global_piqa_nonparallel_generation_srp_latn__dcdf0a40__s000__tb131072__v000",
"uid": "global_piqa_nonparallel_generation_srp_latn__dcdf0a40__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_swe_latn__123d631c__s000__tb131072__v001\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s000__tb131072__v001",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s000__tb131072__v001",
"axis": "code",
"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_swe_latn__123d631c__s000__tb262144__v002\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s000__tb262144__v002",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s000__tb262144__v002",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s000__tb65536__v000",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s000__tb65536__v000",
"axis": "code",
"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_swe_latn__123d631c__s001__tb131072__v004\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s001__tb131072__v004",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s001__tb131072__v004",
"axis": "code",
"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_swe_latn__123d631c__s001__tb262144__v005\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s001__tb262144__v005",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s001__tb262144__v005",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s001__tb65536__v003\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c_... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s001__tb65536__v003",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s001__tb65536__v003",
"axis": "code",
"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_swe_latn__123d631c__s002__tb131072__v007\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s0... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s002__tb131072__v007",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s002__tb131072__v007",
"axis": "code",
"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_swe_latn__123d631c__s002__tb262144__v008\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s0... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s002__tb262144__v008",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s002__tb262144__v008",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s002__tb65536__v006\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s002... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s002__tb65536__v006",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s002__tb65536__v006",
"axis": "code",
"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_swe_latn__123d631c__s003__tb131072__v010\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s003__tb131072__v010",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s003__tb131072__v010",
"axis": "code",
"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_swe_latn__123d631c__s003__tb262144__v011\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s003__tb262144__v011",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s003__tb262144__v011",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s003__tb65536__v009\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s003__tb65536__v009",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s003__tb65536__v009",
"axis": "code",
"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_swe_latn__123d631c__s004__tb131072__v013\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4924.861116211654, \"nll_base\": 8.50205135345459, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s004__tb131072__v013",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s004__tb131072__v013",
"axis": "code",
"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_swe_latn__123d631c__s004__tb262144__v014\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4924.861116211654, \"nll_base\": 8.50205135345459, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s004__tb262144__v014",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s004__tb262144__v014",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s004__tb65536__v012\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4924.861116211654, \"nll_base\": 8.50205135345459, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c_... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s004__tb65536__v012",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s004__tb65536__v012",
"axis": "code",
"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_swe_latn__123d631c__s005__tb131072__v016\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s005__tb131072__v016",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s005__tb131072__v016",
"axis": "code",
"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_swe_latn__123d631c__s005__tb262144__v017\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s005__tb262144__v017",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s005__tb262144__v017",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s005__tb65536__v015\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s005__tb65536__v015",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s005__tb65536__v015",
"axis": "code",
"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_swe_latn__123d631c__s006__tb131072__v019\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s006__tb131072__v019",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s006__tb131072__v019",
"axis": "code",
"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_swe_latn__123d631c__s006__tb262144__v020\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s006__tb262144__v020",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s006__tb262144__v020",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s006__tb65536__v018\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c_... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s006__tb65536__v018",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s006__tb65536__v018",
"axis": "code",
"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_swe_latn__123d631c__s007__tb131072__v022\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s0... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s007__tb131072__v022",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s007__tb131072__v022",
"axis": "code",
"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_swe_latn__123d631c__s007__tb262144__v023\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s0... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s007__tb262144__v023",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s007__tb262144__v023",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s007__tb65536__v021\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s007... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s007__tb65536__v021",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s007__tb65536__v021",
"axis": "code",
"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_swe_latn__123d631c__s008__tb131072__v025\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s008__tb131072__v025",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s008__tb131072__v025",
"axis": "code",
"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_swe_latn__123d631c__s008__tb262144__v026\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s008__tb262144__v026",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s008__tb262144__v026",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s008__tb65536__v024\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s008__tb65536__v024",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s008__tb65536__v024",
"axis": "code",
"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_swe_latn__123d631c__s009__tb131072__v028\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4924.861116211654, \"nll_base\": 8.50205135345459, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s009__tb131072__v028",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s009__tb131072__v028",
"axis": "code",
"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_swe_latn__123d631c__s009__tb262144__v029\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4924.861116211654, \"nll_base\": 8.50205135345459, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s009__tb262144__v029",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s009__tb262144__v029",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s009__tb65536__v027\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4924.861116211654, \"nll_base\": 8.50205135345459, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c_... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s009__tb65536__v027",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s009__tb65536__v027",
"axis": "code",
"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_swe_latn__123d631c__s010__tb131072__v031\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s010__tb131072__v031",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s010__tb131072__v031",
"axis": "code",
"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_swe_latn__123d631c__s010__tb262144__v032\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s010__tb262144__v032",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s010__tb262144__v032",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s010__tb65536__v030\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4442.412932025673, \"nll_base\": 8.398952960968018, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s010__tb65536__v030",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s010__tb65536__v030",
"axis": "code",
"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_swe_latn__123d631c__s011__tb131072__v034\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s011__tb131072__v034",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s011__tb131072__v034",
"axis": "code",
"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_swe_latn__123d631c__s011__tb262144__v035\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s011__tb262144__v035",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s011__tb262144__v035",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s011__tb65536__v033\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 5169.900051663273, \"nll_base\": 8.55060863494873, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c_... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s011__tb65536__v033",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s011__tb65536__v033",
"axis": "code",
"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_swe_latn__123d631c__s012__tb131072__v037\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s0... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s012__tb131072__v037",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s012__tb131072__v037",
"axis": "code",
"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_swe_latn__123d631c__s012__tb262144__v038\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s0... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s012__tb262144__v038",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s012__tb262144__v038",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s012__tb65536__v036\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4483.696626445966, \"nll_base\": 8.408203125, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c__s012... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s012__tb65536__v036",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s012__tb65536__v036",
"axis": "code",
"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_swe_latn__123d631c__s013__tb131072__v040\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s013__tb131072__v040",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s013__tb131072__v040",
"axis": "code",
"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_swe_latn__123d631c__s013__tb262144__v041\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d63... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s013__tb262144__v041",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s013__tb262144__v041",
"axis": "code",
"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_piqa_nonparallel_generation_swe_latn__123d631c__s013__tb65536__v039\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4924.198924122767, \"nll_base\": 8.501916885375977, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s013__tb65536__v039",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s013__tb65536__v039",
"axis": "code",
"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_swe_latn__123d631c__s014__tb65536__v042\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4924.861116211654, \"nll_base\": 8.50205135345459, \"eval_ref\": \"global_piqa_nonparallel_generation_swe_latn__123d631c_... | {
"index": "global_piqa_nonparallel_generation_swe_latn__123d631c__s014__tb65536__v042",
"uid": "global_piqa_nonparallel_generation_swe_latn__123d631c__s014__tb65536__v042",
"axis": "code",
"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_tam_latn__0ee9697d__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3555.659211637352, \"nll_base\": 8.176295757293701, \"eval_ref\": \"global_piqa_nonparallel_generation_tam_latn__0ee969... | {
"index": "global_piqa_nonparallel_generation_tam_latn__0ee9697d__s000__tb262144__v000",
"uid": "global_piqa_nonparallel_generation_tam_latn__0ee9697d__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_piqa_nonparallel_generation_yue_hant__59ed7d44__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4152.7272140577215, \"nll_base\": 8.331520557403564, \"eval_ref\": \"global_piqa_nonparallel_generation_yue_hant__59ed7d4... | {
"index": "global_piqa_nonparallel_generation_yue_hant__59ed7d44__s000__tb65536__v000",
"uid": "global_piqa_nonparallel_generation_yue_hant__59ed7d44__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_parallel_cloze_ell_grek__0cf6479b__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 1.8878461075256912, \"nll_base\": 0.6354365535566763, \"eval_ref\": \"global_piqa_parallel_cloze_ell_grek__0cf6479b__s000__tb6553... | {
"index": "global_piqa_parallel_cloze_ell_grek__0cf6479b__s000__tb65536__v000",
"uid": "global_piqa_parallel_cloze_ell_grek__0cf6479b__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_parallel_cloze_pol_latn__4f39a6e9__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3.341988624862289, \"nll_base\": 1.2065660264756946, \"eval_ref\": \"global_piqa_parallel_cloze_pol_latn__4f39a6e9__s000__tb262... | {
"index": "global_piqa_parallel_cloze_pol_latn__4f39a6e9__s000__tb262144__v000",
"uid": "global_piqa_parallel_cloze_pol_latn__4f39a6e9__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_piqa_parallel_cloze_spa_latn_mexi__f8952d80__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 5.084985502710187, \"nll_base\": 1.6262921785053455, \"eval_ref\": \"global_piqa_parallel_cloze_spa_latn_mexi__f8952d80__s... | {
"index": "global_piqa_parallel_cloze_spa_latn_mexi__f8952d80__s000__tb262144__v000",
"uid": "global_piqa_parallel_cloze_spa_latn_mexi__f8952d80__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_piqa_parallel_cloze_swe_latn__f1a308cf__s001__tb131072__v005\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4.191752541626233, \"nll_base\": 1.4331189141757246, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s001__tb131... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s001__tb131072__v005",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s001__tb131072__v005",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s001__tb65536__v004\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4.191752541626233, \"nll_base\": 1.4331189141757246, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s001__tb65536... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s001__tb65536__v004",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s001__tb65536__v004",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s002__tb131072__v008\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3.375337703890165, \"nll_base\": 1.2164953797306848, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s002__tb131... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s002__tb131072__v008",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s002__tb131072__v008",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s002__tb65536__v007\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3.375337703890165, \"nll_base\": 1.2164953797306848, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s002__tb65536... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s002__tb65536__v007",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s002__tb65536__v007",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s003__tb65536__v010\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3.930276031265192, \"nll_base\": 1.368709660377823, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s003__tb65536_... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s003__tb65536__v010",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s003__tb65536__v010",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s004__tb131072__v014\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3.839188416187537, \"nll_base\": 1.3452609943109248, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s004__tb131... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s004__tb131072__v014",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s004__tb131072__v014",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s006__tb131072__v020\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4.0945369367404725, \"nll_base\": 1.4096536307499326, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s006__tb13... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s006__tb131072__v020",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s006__tb131072__v020",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s006__tb65536__v019\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4.0945369367404725, \"nll_base\": 1.4096536307499326, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s006__tb6553... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s006__tb65536__v019",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s006__tb65536__v019",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s007__tb65536__v022\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3.545707836077202, \"nll_base\": 1.265737811724345, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s007__tb65536_... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s007__tb65536__v022",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s007__tb65536__v022",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s009__tb65536__v028\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3.9803857150043522, \"nll_base\": 1.3813787279182306, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s009__tb6553... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s009__tb65536__v028",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s009__tb65536__v028",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s010__tb65536__v031\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3.9660718372862362, \"nll_base\": 1.3777761432180378, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s010__tb6553... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s010__tb65536__v031",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s010__tb65536__v031",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s011__tb65536__v034\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4.510834955324613, \"nll_base\": 1.5064822705976801, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s011__tb65536... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s011__tb65536__v034",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s011__tb65536__v034",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s012__tb131072__v038\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3.3847994074050174, \"nll_base\": 1.2192946453483737, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s012__tb13... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s012__tb131072__v038",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s012__tb131072__v038",
"axis": "code",
"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_parallel_cloze_swe_latn__f1a308cf__s012__tb65536__v037\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3.3847994074050174, \"nll_base\": 1.2192946453483737, \"eval_ref\": \"global_piqa_parallel_cloze_swe_latn__f1a308cf__s012__tb6553... | {
"index": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s012__tb65536__v037",
"uid": "global_piqa_parallel_cloze_swe_latn__f1a308cf__s012__tb65536__v037",
"axis": "code",
"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_parallel_cloze_swh_latn__da71424c__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4.103335458537533, \"nll_base\": 1.4118001694495317, \"eval_ref\": \"global_piqa_parallel_cloze_swh_latn__da71424c__s000__tb65536... | {
"index": "global_piqa_parallel_cloze_swh_latn__da71424c__s000__tb65536__v000",
"uid": "global_piqa_parallel_cloze_swh_latn__da71424c__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_parallel_generation_amh_ethi__dbb50279__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 4020.633727215864, \"nll_base\": 8.299194812774658, \"eval_ref\": \"global_piqa_parallel_generation_amh_ethi__dbb50279__s0... | {
"index": "global_piqa_parallel_generation_amh_ethi__dbb50279__s000__tb131072__v000",
"uid": "global_piqa_parallel_generation_amh_ethi__dbb50279__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_parallel_generation_arb_arab__f4f1d602__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2288.9018635894554, \"nll_base\": 7.735827445983887, \"eval_ref\": \"global_piqa_parallel_generation_arb_arab__f4f1d602__s00... | {
"index": "global_piqa_parallel_generation_arb_arab__f4f1d602__s000__tb65536__v000",
"uid": "global_piqa_parallel_generation_arb_arab__f4f1d602__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_parallel_generation_ben_latn__eedac295__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2668.228824948741, \"nll_base\": 7.889170169830322, \"eval_ref\": \"global_piqa_parallel_generation_ben_latn__eedac295__s000... | {
"index": "global_piqa_parallel_generation_ben_latn__eedac295__s000__tb65536__v000",
"uid": "global_piqa_parallel_generation_ben_latn__eedac295__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_parallel_generation_bxr_cyrl__db84ac2a__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 4015.317050085819, \"nll_base\": 8.297871589660645, \"eval_ref\": \"global_piqa_parallel_generation_bxr_cyrl__db84ac2a__s000... | {
"index": "global_piqa_parallel_generation_bxr_cyrl__db84ac2a__s000__tb65536__v000",
"uid": "global_piqa_parallel_generation_bxr_cyrl__db84ac2a__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_parallel_generation_kan_latn__df9990bc__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2655.0567690413072, \"nll_base\": 7.884221315383911, \"eval_ref\": \"global_piqa_parallel_generation_kan_latn__df9990bc__s00... | {
"index": "global_piqa_parallel_generation_kan_latn__df9990bc__s000__tb65536__v000",
"uid": "global_piqa_parallel_generation_kan_latn__df9990bc__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_parallel_generation_mal_mlym__a3a1380e__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2166.592261390729, \"nll_base\": 7.68091082572937, \"eval_ref\": \"global_piqa_parallel_generation_mal_mlym__a3a1380e__s00... | {
"index": "global_piqa_parallel_generation_mal_mlym__a3a1380e__s000__tb131072__v000",
"uid": "global_piqa_parallel_generation_mal_mlym__a3a1380e__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_parallel_generation_nld_latn__1c61f8d3__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2565.551174783635, \"nll_base\": 7.849928617477417, \"eval_ref\": \"global_piqa_parallel_generation_nld_latn__1c61f8d3__s0... | {
"index": "global_piqa_parallel_generation_nld_latn__1c61f8d3__s000__tb131072__v000",
"uid": "global_piqa_parallel_generation_nld_latn__1c61f8d3__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_parallel_generation_pan_guru__42568bf9__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2608.5971101338246, \"nll_base\": 7.866567850112915, \"eval_ref\": \"global_piqa_parallel_generation_pan_guru__42568bf9__s... | {
"index": "global_piqa_parallel_generation_pan_guru__42568bf9__s000__tb131072__v000",
"uid": "global_piqa_parallel_generation_pan_guru__42568bf9__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_parallel_generation_pcm_latn__2cd598c9__s000__tb65536__v000\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 2597.500527150202, \"nll_base\": 7.862304925918579, \"eval_ref\": \"global_piqa_parallel_generation_pcm_latn__2cd598c9__s000... | {
"index": "global_piqa_parallel_generation_pcm_latn__2cd598c9__s000__tb65536__v000",
"uid": "global_piqa_parallel_generation_pcm_latn__2cd598c9__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_parallel_generation_pes_arab__e0bd58c5__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1745.3160290135322, \"nll_base\": 7.464690923690796, \"eval_ref\": \"global_piqa_parallel_generation_pes_arab__e0bd58c5__s... | {
"index": "global_piqa_parallel_generation_pes_arab__e0bd58c5__s000__tb262144__v000",
"uid": "global_piqa_parallel_generation_pes_arab__e0bd58c5__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_piqa_parallel_generation_pol_latn__310d9ce8__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 2381.494732139578, \"nll_base\": 7.77548360824585, \"eval_ref\": \"global_piqa_parallel_generation_pol_latn__310d9ce8__s00... | {
"index": "global_piqa_parallel_generation_pol_latn__310d9ce8__s000__tb262144__v000",
"uid": "global_piqa_parallel_generation_pol_latn__310d9ce8__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_piqa_parallel_generation_slv_latn__adc68bb4__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 2453.7850263544556, \"nll_base\": 7.805387020111084, \"eval_ref\": \"global_piqa_parallel_generation_slv_latn__adc68bb4__s... | {
"index": "global_piqa_parallel_generation_slv_latn__adc68bb4__s000__tb262144__v000",
"uid": "global_piqa_parallel_generation_slv_latn__adc68bb4__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_piqa_parallel_generation_snd_arab__35600297__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 1927.9132650809904, \"nll_base\": 7.564193487167358, \"eval_ref\": \"global_piqa_parallel_generation_snd_arab__35600297__s... | {
"index": "global_piqa_parallel_generation_snd_arab__35600297__s000__tb262144__v000",
"uid": "global_piqa_parallel_generation_snd_arab__35600297__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_piqa_parallel_generation_srp_latn_torl__6e3a788b__s000__tb131072__v000\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 2115.385692670305, \"nll_base\": 7.656992435455322, \"eval_ref\": \"global_piqa_parallel_generation_srp_latn_torl__6e... | {
"index": "global_piqa_parallel_generation_srp_latn_torl__6e3a788b__s000__tb131072__v000",
"uid": "global_piqa_parallel_generation_srp_latn_torl__6e3a788b__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_parallel_generation_swe_latn__d9d0dde4__s000__tb131072__v002\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3250.9726236238057, \"nll_base\": 8.08670949935913, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s0... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s000__tb131072__v002",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s000__tb131072__v002",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s000__tb262144__v000\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3250.9726236238057, \"nll_base\": 8.08670949935913, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s0... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s000__tb262144__v000",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s000__tb262144__v000",
"axis": "code",
"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_piqa_parallel_generation_swe_latn__d9d0dde4__s000__tb65536__v001\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3250.9726236238057, \"nll_base\": 8.08670949935913, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s000... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s000__tb65536__v001",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s000__tb65536__v001",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s001__tb131072__v005\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3678.17570665872, \"nll_base\": 8.210172176361084, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s00... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s001__tb131072__v005",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s001__tb131072__v005",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s001__tb262144__v003\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3678.17570665872, \"nll_base\": 8.210172176361084, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s00... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s001__tb262144__v003",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s001__tb262144__v003",
"axis": "code",
"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_piqa_parallel_generation_swe_latn__d9d0dde4__s001__tb65536__v004\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3678.17570665872, \"nll_base\": 8.210172176361084, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s001_... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s001__tb65536__v004",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s001__tb65536__v004",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s002__tb131072__v008\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3580.0999568344673, \"nll_base\": 8.183145999908447, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s002__tb131072__v008",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s002__tb131072__v008",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s002__tb262144__v006\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3580.0999568344673, \"nll_base\": 8.183145999908447, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s002__tb262144__v006",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s002__tb262144__v006",
"axis": "code",
"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_piqa_parallel_generation_swe_latn__d9d0dde4__s002__tb65536__v007\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3580.0999568344673, \"nll_base\": 8.183145999908447, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s00... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s002__tb65536__v007",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s002__tb65536__v007",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s003__tb131072__v011\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3400.0407051913994, \"nll_base\": 8.131542682647705, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s003__tb131072__v011",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s003__tb131072__v011",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s003__tb262144__v009\", \"budget\": 2048, \"token_budget\": 262144, \"ppl_base\": 3400.0407051913994, \"nll_base\": 8.131542682647705, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s003__tb262144__v009",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s003__tb262144__v009",
"axis": "code",
"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_piqa_parallel_generation_swe_latn__d9d0dde4__s003__tb65536__v010\", \"budget\": 2048, \"token_budget\": 65536, \"ppl_base\": 3400.0407051913994, \"nll_base\": 8.131542682647705, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s00... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s003__tb65536__v010",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s003__tb65536__v010",
"axis": "code",
"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_parallel_generation_swe_latn__d9d0dde4__s004__tb131072__v014\", \"budget\": 2048, \"token_budget\": 131072, \"ppl_base\": 3702.1926333986967, \"nll_base\": 8.216680526733398, \"eval_ref\": \"global_piqa_parallel_generation_swe_latn__d9d0dde4__s... | {
"index": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s004__tb131072__v014",
"uid": "global_piqa_parallel_generation_swe_latn__d9d0dde4__s004__tb131072__v014",
"axis": "code",
"budget": 2048,
"token_budget": 131072,
"variant": "filter"
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.