Summarize harnesses the power of the GPT class of large language models from OpenAI, including GPT-3.5 and GPT-4. These models excel at generating a wide array of content using language and text. In this case, we use this technology to generate contextually relevant summaries expressed in coherent human language
Summarize is our first Generative AI feature, employing a large language model to create intuitive, readable outputs around unstructured open-ended data. Unlike previous AI features, these new natural language summaries require minimal user interpretation, and do not revolve around specific numbers or labels. As AI is a powerful yet assistive tool, we do recommend review of outputs and underlying data.
Use of large language models have been prohibitive in the past due to practices around use of data to train or tune models. These policies have since changed, and data is only used to create outputs, and not for training or tuning models.
Data privacy, security, and ownership are huge concerns with the rapid innovation of AI technology because it relies on massive amounts of data. At Remesh, we take this very seriously, which is why we haven’t adopted LLMs until the recent removal of the requirement to keep and use customers’ input data.
The recent requirement changes mean that the use of data for the Summarize feature conforms to standard Remesh policies, with OpenAI as a Remesh data subprocessor.
Please view our Privacy Policy and Terms of Service. Data is not used to train or tune any models. Please also review OpenAI’s Terms & Policies and specific API Data Usage Policies.
Summarize has a lot of flexibility to meet your research needs. You can use Summarize for responses to individual questions. You can also hand-pick and select multiple questions. And you can even use Summarize to generate a text summary of the entire Conversation!
Want to use Summarize to better understand audience segments? You can do that too! Just select the segments you want to include and generate your summaries.
Summarize optimizes for both performance and relevance. The Remesh platform empowers participants to vote on others’ responses, and Summarize adapts to this audience preference by focusing on key representative responses. This unique process ensures that researchers understand not only how individuals feel, but how they align to ideas collectively.
With Summarize, your research efforts become even more powerful by enabling the most qualified people to tell you which data is the most important: your participants. While conventional tools simply copy and paste responses to generate text outputs, Remesh takes an extra step so that our users can get to insights faster and be more confident in the quality of the outcomes they deliver.