Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code: DatasetGenerationError
Exception: CastError
Message: Couldn't cast
answer_count: int64
answer_options: list<item: list<item: string>>
child 0, item: list<item: string>
child 0, item: string
answers: list<item: string>
child 0, item: string
context: string
context_id: string
context_sha256: string
eval_type: string
id: string
problem_key: string
query: string
query_sha256: string
source: string
source_index: int64
split: string
task_lang: string
task_type: string
work_lang: string
year: int64
metadata: null
messages: list<item: struct<content: string, role: string>>
child 0, item: struct<content: string, role: string>
child 0, content: string
child 1, role: string
to
{'messages': List({'content': Value('string'), 'role': Value('string')}), 'metadata': Json(decode=True)}
because column names don't match
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1816, in _prepare_split_single
for key, table in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
answer_count: int64
answer_options: list<item: list<item: string>>
child 0, item: list<item: string>
child 0, item: string
answers: list<item: string>
child 0, item: string
context: string
context_id: string
context_sha256: string
eval_type: string
id: string
problem_key: string
query: string
query_sha256: string
source: string
source_index: int64
split: string
task_lang: string
task_type: string
work_lang: string
year: int64
metadata: null
messages: list<item: struct<content: string, role: string>>
child 0, item: struct<content: string, role: string>
child 0, content: string
child 1, role: string
to
{'messages': List({'content': Value('string'), 'role': Value('string')}), 'metadata': Json(decode=True)}
because column names don't match
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1869, in _prepare_split_single
raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the datasetNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
messages list | metadata unknown |
|---|---|
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201201",
"source_ids": [
"012012010200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201504",
"source_ids": [
"012015040200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201005",
"source_ids": [
"012010050402"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200405:f412ecaee90b9982",
"problem_key": "01200405",
"source_ids": [
"012004050100"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201504:86e2566be3d8fc6d",
"problem_key": "01201504",
"source_ids": [
"012015040200"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201103",
"source_ids": [
"012011030100"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2024-individual-1",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2024-individual-p1-sub-b-atomic-b.4"
],
"split": "train",
"trace_provenance": "manual_expansion_v12"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201002",
"source_ids": [
"012010020300"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01201405:db73d57dd6b4181c",
"problem_key": "01201405",
"source_ids": [
"012014050100",
"012014050200",
"012014050300"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201302",
"source_ids": [
"012013020100"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2016-individual-5",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2016-individual-p5-sub-b-atomic-b.6"
],
"split": "train",
"trace_provenance": "answer_trusted_facts"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200702",
"source_ids": [
"012007020200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200502",
"source_ids": [
"012005020200"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01201702:84a86cb71ccfdf1d",
"problem_key": "01201702",
"source_ids": [
"012017020100"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201005:b3af66a248621373",
"problem_key": "01201005",
"source_ids": [
"012010050402"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200804",
"source_ids": [
"012008040200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201605",
"source_ids": [
"012016050100"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200801:7def61d379bc28a5",
"problem_key": "01200801",
"source_ids": [
"012008010100"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201502",
"source_ids": [
"012015020100"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201204:d58f1d54640ab15f",
"problem_key": "01201204",
"source_ids": [
"012012040300"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2018-individual-1",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2018-individual-p1-sub-a-atomic-a.6"
],
"split": "train",
"trace_provenance": "manual_expansion_v17"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201102:9d545e6f33802ed8",
"problem_key": "01201102",
"source_ids": [
"012011020100"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2006-individual-2",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2006-individual-p2-sub-b-atomic-b.1"
],
"split": "train",
"trace_provenance": "manual_expansion_v14"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201104",
"source_ids": [
"012011040300"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201002",
"source_ids": [
"012010020300"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200901",
"source_ids": [
"012009010200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01200405:f412ecaee90b9982",
"problem_key": "01200405",
"source_ids": [
"012004050100"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2006-individual-2",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2006-individual-p2-sub-a-atomic-a.12"
],
"split": "train",
"trace_provenance": "manual_expansion_v14"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01200902:a1c24cf560af260a",
"problem_key": "01200902",
"source_ids": [
"012009020100"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200805",
"source_ids": [
"012008050200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2019-individual-5",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2019-individual-p5-sub-c-atomic-c.4"
],
"split": "train",
"trace_provenance": "answer_trusted_facts"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200804",
"source_ids": [
"012008040200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200502",
"source_ids": [
"012005020100"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200303",
"source_ids": [
"012003030200"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2017-individual-3",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2017-individual-p3-sub-a-atomic-a.1"
],
"split": "train",
"trace_provenance": "manual_expansion_v14"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200801",
"source_ids": [
"012008010200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2017-individual-2",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2017-individual-p2-sub-c-atomic-c.6"
],
"split": "train",
"trace_provenance": "manual_expansion_v13"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2019-individual-5",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2019-individual-p5-sub-a-atomic-a.3"
],
"split": "train",
"trace_provenance": "answer_trusted_facts"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201204",
"source_ids": [
"012012040100"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2008-individual-1",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2008-individual-p1-sub-b-atomic-b.3"
],
"split": "train",
"trace_provenance": "manual_expansion_v16"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200803",
"source_ids": [
"012008030300"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01201404:08c94556a6ef2cc3",
"problem_key": "01201404",
"source_ids": [
"012014040100"
],
"split": "train",
"teacher": "codex-b"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01200905:809e899db2bcef81",
"problem_key": "01200905",
"source_ids": [
"012009050100",
"012009050200"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201002:f1ed7d934eb3cf71",
"problem_key": "01201002",
"source_ids": [
"012010020200"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200401",
"source_ids": [
"012004010100"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200303:aa9c60484df5499f",
"problem_key": "01200303",
"source_ids": [
"012003030200"
],
"split": "train",
"teacher": "codex-b"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200805",
"source_ids": [
"012008050200"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201203",
"source_ids": [
"012012030200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201204",
"source_ids": [
"012012040200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200604",
"source_ids": [
"012006040100"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200803",
"source_ids": [
"012008030202"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2019-individual-5",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2019-individual-p5-sub-a-atomic-a.4"
],
"split": "train",
"trace_provenance": "answer_trusted_facts"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201301:5ac9e726319bda01",
"problem_key": "01201301",
"source_ids": [
"012013010200"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2018-individual-2",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2018-individual-p2-sub-b-atomic-b.4"
],
"split": "train",
"trace_provenance": "manual_expansion_v14"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2018-individual-1",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2018-individual-p1-sub-a-atomic-a.4"
],
"split": "train",
"trace_provenance": "manual_expansion_v17"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201502:9e02e466ac07de84",
"problem_key": "01201502",
"source_ids": [
"012015020100"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200702:b0a09dd9a6477edd",
"problem_key": "01200702",
"source_ids": [
"012007020200"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2018-individual-1",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2018-individual-p1-sub-a-atomic-a.11"
],
"split": "train",
"trace_provenance": "manual_expansion_v17"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200801",
"source_ids": [
"012008010100"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201603",
"source_ids": [
"012016030200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200601",
"source_ids": [
"012006010300"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200404",
"source_ids": [
"012004040200"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201302:16fde563fd7d2004",
"problem_key": "01201302",
"source_ids": [
"012013020100"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200502",
"source_ids": [
"012005020200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201201",
"source_ids": [
"012012010300"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201401:c7d4c553167b8c40",
"problem_key": "01201401",
"source_ids": [
"012014010100"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200304",
"source_ids": [
"012003040400"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201703",
"source_ids": [
"012017030200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201205",
"source_ids": [
"012012050200"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200504",
"source_ids": [
"012005040200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201103",
"source_ids": [
"012011030100"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2018-individual-3",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2018-individual-p3-sub-a-atomic-a.9"
],
"split": "train",
"trace_provenance": "manual_expansion_v14"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200401",
"source_ids": [
"012004010200"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200801",
"source_ids": [
"012008010100"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01200601:d079fd2a8c18c1cc",
"problem_key": "01200601",
"source_ids": [
"012006010100",
"012006010200",
"012006010300"
],
"split": "train",
"teacher": "codex-b"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2017-individual-2",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2017-individual-p2-sub-b-atomic-b.1"
],
"split": "train",
"trace_provenance": "manual_expansion_v13"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201102",
"source_ids": [
"012011020100"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2018-individual-1",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2018-individual-p1-sub-a-atomic-a.2"
],
"split": "train",
"trace_provenance": "manual_expansion_v17"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200601:d079fd2a8c18c1cc",
"problem_key": "01200601",
"source_ids": [
"012006010300"
],
"split": "train",
"teacher": "codex-b"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2006-individual-2",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2006-individual-p2-sub-a-atomic-a.7"
],
"split": "train",
"trace_provenance": "manual_expansion_v14"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201605",
"source_ids": [
"012016050200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01201005:b3af66a248621373",
"problem_key": "01201005",
"source_ids": [
"012010050401"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In DIRECT mode output only FINAL ANSWERS followed by exactly the requested number of answer lines, with no ... | {
"example_type": "direct_aux",
"problem_key": "ioling:2019-individual-5",
"source": "agurung/ioling:sft_selected_student_traces_v9_expanded_clean",
"source_ids": [
"iol-2019-individual-p5-sub-c-atomic-c.1"
],
"split": "train",
"trace_provenance": "answer_trusted_facts"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "induce",
"packet_id": "train:01201705:6724441bf700ee70",
"problem_key": "01201705",
"source_ids": [
"012017050100",
"012017050200",
"012017050300"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01201203",
"source_ids": [
"012012030300"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200905",
"source_ids": [
"012009050200"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200504:d5addd9385c50dcc",
"problem_key": "01200504",
"source_ids": [
"012005040200"
],
"split": "train",
"teacher": "codex-a"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01201705",
"source_ids": [
"012017050200"
],
"split": "train",
"variant": 1
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200804:b8e5dda8681c266a",
"problem_key": "01200804",
"source_ids": [
"012008040100"
],
"split": "train",
"teacher": "codex-c"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "format_repair",
"problem_key": "01200901",
"source_ids": [
"012009010100"
],
"split": "train",
"variant": 0
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "direct",
"problem_key": "01200803",
"source_ids": [
"012008030201"
],
"split": "train"
} |
[
{
"content": "You solve unseen linguistics-olympiad problems from the evidence in each prompt. Obey the requested mode exactly and preserve every unfamiliar character, accent, space, and punctuation mark. In INDUCE mode output only a compact reusable RULE SHEET and do not answer the queries. In APPLY, DIRECT, o... | {
"example_type": "apply",
"packet_id": "train:01200701:00b45f7ff3036c92",
"problem_key": "01200701",
"source_ids": [
"012007010100"
],
"split": "train",
"teacher": "codex-b"
} |
End of preview.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
First Qwen3.5-9B SFT
This is the deliberately small first experiment: train on verified historical IOL material, measure on untouched development problems, then run the resulting checkpoint on the locked 2021–2023 Linguini test and the competition.
Data
final/train.jsonl: 616 chat examples (including three marked repeats so all records fit complete batches of eight).final/eval.jsonl: 61 examples from ten held-out pre-2021 problems.locked/locked_test.jsonl: 32 rows from fourteen 2021–2023 problems; never passed to Tinker training or loss evaluation.- The training mix contains
INDUCE,APPLY,DIRECT, format-repair, and 143 manually approved auxiliary DIRECT traces fromagurung/iolingafter all 2021–2023 sources were removed.
The split is by complete olympiad problem ID, not by row. The contamination gate additionally checks normalized exact text, long substrings, Unicode and typography variants, IOL identifiers, and fuzzy character/word n-grams.
Rebuild and verify
uv run --with pyarrow --with pandas python scripts/prepare_linguini_splits.py
uv run --with pyarrow --with pandas python scripts/prepare_ioling_aux.py
python3 scripts/build_sft_examples.py
python3 scripts/check_contamination.py \
--locked sft_data/locked/locked_test.jsonl \
--lock-manifest sft_data/locked/lock_manifest.json \
--train sft_data/final/train.jsonl \
--report sft_data/reports/contamination_report.json \
--fail-on-review
rule_sheets.jsonl is already checked in; rebuilding it requires rerunning the
teacher-agent step, not Tinker.
Run on Tinker
First run the exact renderer/token/cost preflight. It does not use credits:
(when running from a clone of the public dataset repo, replace
sft_data/train_tinker.py with train_tinker.py).
uv run \
--with 'tinker>=0.22.3' \
--with 'tinker-cookbook @ git+https://github.com/thinking-machines-lab/tinker-cookbook.git@3e04119ce293a2b6ba5284e35267c9ba6d27c5da' \
python sft_data/train_tinker.py --preflight-only
Then set your key and run the same command without --preflight-only:
export TINKER_API_KEY='your-key'
uv run \
--with 'tinker>=0.22.3' \
--with 'tinker-cookbook @ git+https://github.com/thinking-machines-lab/tinker-cookbook.git@3e04119ce293a2b6ba5284e35267c9ba6d27c5da' \
python sft_data/train_tinker.py
The fixed first run is Qwen3.5-9B, non-thinking renderer, LoRA rank 32,
assistant-only loss, learning rate 1e-4 with linear decay, batch size 8, and
two epochs. A checkpoint and NLL evaluation are produced at each epoch.
The verified final preflight is 334,266 train tokens per epoch, with longest train/eval records of 1,510/1,743 tokens (well below 4,096). At the current official rate of $1.463 per million training tokens, two epochs cost $0.98; budgeting $2 leaves comfortable room for the two eval passes.
- Downloads last month
- 19