The dataset viewer is not available for this split.
Unexpected token '<', "<html>
<h"... is not valid JSON
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Subsets and Splits
Interquartile Response Length Filter
Identifies high-quality responses within each category by filtering to the interquartile range of token lengths, helping find appropriately detailed answers that avoid being too brief or overly verbose.
Open QA Response Token Count Analysis
Identifies and retrieves responses within the typical token length range (25th to 75th percentile), revealing the central tendency of response lengths in open QA tasks and helping understand the standard complexity of model responses.
Information Extraction Entries
Retrieves all training examples labeled as 'information_extraction' which provides basic filtering but doesn't reveal meaningful patterns or relationships in the data.
Count Information Extraction Entries
Counts the number of training examples in the information extraction category, providing basic dataset balance information but offering limited analytical insight.
Category Samples Summary
Randomly samples 100 examples from each of four task categories to create a balanced subset, which helps in fair comparison across different NLP tasks but doesn't reveal deeper patterns or relationships in the data.
Travel-related Queries
Retrieves all records containing travel-related keywords in the instruction, context, or response, providing a basic filter for travel-related data.
Sample Prompts and Completions
Retrieves a sample of 12,000 rows showing prompts and their corresponding completions, providing a basic overview of the dataset's content.
Sample Prompts and Completions
Retrieves a sample of 4000 prompts and their corresponding completions, providing a basic overview of the dataset's content.
Machine Learning Instructions & Responses
Retrieves a limited set of entries where the instruction or response contains the phrase "machine learning", providing basic filtering of relevant data.
Short Responses Filter
The query filters responses to those under 2048 characters in length, providing basic filtering with limited analytical value.