Copying and Exporting Texts
After PLAI processes the prompt, you can copy the proposed text in its entirety or export it as a Doc or PDF. Use the three icons below the text for this purpose.
Selecting LLM (Large Language Model)
PLAI allows you to choose from the best available Large Language Models for your prompts. Currently, you can choose between GPT-4, ClaudeV2, and GPT-4 Turbo for each prompt. Click on the icon for the desired LLM in the left panel before submitting the prompt and activate the checkbox for the chosen LLM. Click on the…
Read more
Considering Chat History and New Sessions
During prompt execution, previous answers and dialog texts can be considered. This is useful if you refer to previous responses in your prompt and want to elaborate on specific points. You can engage in a dialogue, have a discussion, or refer to previous answers in a new task. Activate the “Consider Chat History” checkbox before…
Read more
Creating Collections
Open the collections dialog box through the left-side menu. If there are existing collections, they will be displayed. Click on “Add Collection” and choose a suitable name for your collection, such as “Employment Contracts.” Save it with the floppy disk icon. Use the arrow symbol to load corresponding files into the collection. The upper limit…
Read more
LLM
PLAI provides various Large Language Models that you can choose for your legal work. A Large Language Model or LLM is a computer program trained to understand and generate human-like texts. It can answer questions posed in natural language, compose texts, and be employed in various language-related tasks. It learns from vast amounts of human…
Read more
Temperature
The temperature parameter influences the randomness of the LLM’s response. The outputs are not deterministic; instead, probabilities for various responses are calculated. The temperature, typically ranging from 0 to 1, regulates this probability distribution. A higher temperature value (close to 1) increases randomness, leading to more diverse and creative responses. A lower value (close to…
Read more
