reasoning-sft-OpenThoughts3-1.2M-450K

Converted version of open-thoughts/OpenThoughts3-1.2M, filtered to rows with exactly one valid <think>...</think> block. 750K rows were dropped due to missing or malformed think tags.

Format

Each row has three columns:

• input — list of dicts (conversation turns with role and content; human → user, last gpt turn removed)
• response — gpt response string including <think> reasoning block
• domain — task domain (math, code, science)

Domain Distribution

Domain	Count	%
math	274,262	60.86%
science	88,474	19.63%
code	87,874	19.50%

Usage

import random
import pyarrow.parquet as pq
from huggingface_hub import hf_hub_download

fpath = hf_hub_download(
    repo_id="AmanPriyanshu/reasoning-sft-OpenThoughts3-1.2M-450K",
    repo_type="dataset",
    filename="data_converted.parquet",
    local_dir="./tmp_openthoughts_peek"
)

pf = pq.ParquetFile(fpath)
rows = {"input": [], "response": [], "domain": []}
for batch in pf.iter_batches(batch_size=65_536):
    d = batch.to_pydict()
    rows["input"].extend(d["input"])
    rows["response"].extend(d["response"])
    rows["domain"].extend(d["domain"])

total = len(rows["input"])
for idx in random.sample(range(total), 3):
    print(f"\n{'='*80}\nRow {idx:,} / {total:,}  |  domain: {rows['domain'][idx]}\n{'='*80}")
    for msg in rows["input"][idx]:
        print(f"\n  [{msg['role']}]\n  {msg['content'][:300]}{'...' if len(msg['content']) > 300 else ''}")
    print(f"\n[response]\n{rows['response'][idx][:600]}{'...' if len(rows['response'][idx]) > 600 else ''}")

License

Apache 2.0

Credits

Original dataset: open-thoughts/OpenThoughts3-1.2M
Annotations generated with QwQ-32B