Built using our deep expertise in GenAI use cases. Powered by exceptional raters. Sure to improve your model performance.
Access Datasets at Scale, with Speed
Strong, high-touch QC processes
High Inter-Rater Reliability
Diverse Demographics
Experience across coding, finance, content moderation, & more
This dataset enables use-cases like Market Research/Analytics over infographics. Each datapoint consists of vivid images (infographics, reports, charts), a Prompt (complicated analytical question over the image), and a detailed step-by-step Response (the answer).
This a high quality, single turn Indic Language LLM fine-tuning dataset which enables general purpose LLMs to extend their multilingual capabilities.
This is a high quality, single shot Text2SQL LLM fine-tuning dataset which enables Conversational Business Intelligence use-cases. Each datapoint consists of a Natural Language Question (NLQ), SQL Query pair. The NLQ addresses a complicated business insight question over the DB, and the SQL query solves it.
This dataset consists of a high quality, single turn Indic Language LLM RLHF dataset which enables general purpose LLMs to extend their multilingual capabilities.
This is a PPO dataset that assembles a high-quality RLHF dataset designed to address end-consumer use cases across various domains. Each data point consists of a challenging prompt along with a pair of responses generated by two different large language models (LLMs).
This dataset utilises financial documents from U.S. companies, including annual reports (Form 10-K), annual general meetings, and quarterly reports. To simulate real-world scenarios, the annotators (finance experts) have carefully designed questions that reflect the needs of different financial companies and varying levels of financial literacy.
This dataset enables use-cases like Market Research/Analytics over infographics. Each datapoint consists of vivid images (infographics, reports, charts), a Prompt (complicated analytical question over the image), and a detailed step-by-step Response (the answer).
This a high quality, single turn Indic Language LLM fine-tuning dataset which enables general purpose LLMs to extend their multilingual capabilities.
This is a high quality, single shot Text2SQL LLM fine-tuning dataset which enables Conversational Business Intelligence use-cases. Each datapoint consists of a Natural Language Question (NLQ), SQL Query pair. The NLQ addresses a complicated business insight question over the DB, and the SQL query solves it.
This dataset consists of a high quality, single turn Indic Language LLM RLHF dataset which enables general purpose LLMs to extend their multilingual capabilities.
This is a PPO dataset that assembles a high-quality RLHF dataset designed to address end-consumer use cases across various domains. Each data point consists of a challenging prompt along with a pair of responses generated by two different large language models (LLMs).
This dataset utilises financial documents from U.S. companies, including annual reports (Form 10-K), annual general meetings, and quarterly reports. To simulate real-world scenarios, the annotators (finance experts) have carefully designed questions that reflect the needs of different financial companies and varying levels of financial literacy.
Fraud-mitigation
Certifications
