Sign in to view source links and access this dataset
Description
A text dataset for instruction tuning of language models, created by kshitijthakkar and last updated in December 2025. The dataset contains at least 66,652 examples, with token statistics indicating a focus on long-context training. The total token count is 277,312,737, with a mean of 4,160 tokens per example.
Use Cases
Fine-tuning language models for instruction-following based on the described instruction-tuning purpose.
Training models to handle long-context inputs based on the high percentage of examples exceeding 1,024 tokens.
Benchmarking model performance on sequences of varying lengths based on the provided token count distribution statistics.
Strengths
All 66,652 examples exceed 512 tokens, providing consistent long-context data.
99.53% of examples exceed 1,024 tokens, indicating a strong focus on long sequences.
Provides detailed token count statistics, including min, max, mean, median, and total tokens.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
huggingface
Freshness
Last updated 2025-12-16 06:14:34; freshness should be verified.
License is unknown; terms of use must be verified before application.