Systems and methods for providing user interfaces to converse with a corpus of electronic documents via a large language model

1 . A system configured for providing user interfaces for users to textually converse with a corpus of electronic documents via a large language model, wherein the large language model has been trained on at least a million documents, wherein the large language model includes a neural network using over a billion parameters and/or weights, the system comprising:

one or more hardware processors configured by machine-readable instructions to:

effectuate a presentation of a first user interface, the first user interface being configured to obtain entry of user input from a user to:

(i) select multiple documents to be provided as input to a large language model for a particular textual conversation between the user and the multiple documents, wherein the multiple documents form the corpus of electronic documents,

(ii) enter queries regarding the multiple documents that have been selected, and

(iii) navigate between a set of different portions of the first user interface, wherein the set of different portions includes:

(a) a first portion configured to select, by the user, an individual textual conversation from a set of textual conversations that are selectable, wherein the set of textual conversations includes the particular textual conversation,

(b) a second portion configured to enter, by the user, an individual query and present, to the user, an individual reply to the individual query,

(c) a third portion configured to select, by the user, an individual document from the multiple documents for presentation in a fourth portion, and

(d) the fourth portion configured to present, to the user, at least part of the individual document as both (1) selected in the third portion and (2) determined to be relevant to the individual reply;

responsive to selection of the individual textual conversation, provide the individual query as a prompt to the large language model;

obtain the individual reply from the large language model, wherein the individual reply is presented in the second portion of the first user interface;

determine the individual document from the multiple documents that is relevant to the individual reply; and

present the individual document in the fourth portion of the first user interface, wherein a segment in the individual document that is relevant to the individual reply is emphasized in the fourth portion of the first user interface.

2 . The system of claim 1 , wherein the second portion is specific to the individual textual conversation from the set of textual conversations that has been selected through the first portion of the first user interface.

3 . The system of claim 1 , wherein the second portion presents modifications to the corpus of electronic documents.

4 . The system of claim 1 , wherein the second portion indicates whether the individual query was scoped to a subset of the corpus of electronic documents.

5 . The system of claim 1 , wherein the large language model is modeled such that the individual reply is limited in scope to one or more statements that have support in the corpus of electronic documents.

6 . The system of claim 1 , wherein the large language model is based on or derived from Generative Pre-trained Transformer 3 (GPT3).

7 . The system of claim 1 , wherein the second portion indicates a subset of the corpus of electronic documents determined to be relevant to the individual reply, wherein the subset includes the individual document.

8 . The system of claim 1 , wherein the individual reply is formatted for presentation based on the individual query.

9 . The system of claim 1 , wherein the segment in the individual document that is relevant to the individual reply is emphasized in the fourth portion of the first user interface by an indicator that indicates provenance of the individual reply.

10 . The system of claim 1 , wherein at least some parts of the first portion, the second portion, the third portion, and the fourth portion are presented to the user at the same time.

11 . A method of providing user interfaces to users to textually converse with a corpus of electronic documents via a large language model, wherein the large language model has been trained on at least a million documents, wherein the large language model includes a neural network using over a billion parameters and/or weights, the method comprising:

effectuating a presentation of a first user interface, the first user interface being configured to obtain entry of user input from a user to

(i) select multiple documents to be provided as input to a large language model for a particular textual conversation between the user and the multiple documents, wherein the multiple documents form the corpus of electronic documents,

(ii) enter queries regarding the multiple documents that have been selected, and

(iii) navigate between a set of different portions of the first user interface, wherein the set of different portions includes

(a) a first portion configured to select, by the user, an individual textual conversation from a set of textual conversations that are selectable, wherein the set of textual conversations includes the particular textual conversation,

(b) a second portion configured to enter, by the user, an individual query and present, to the user, an individual reply to the individual query,

(c) a third portion configured to select, by the user, an individual document from the multiple documents for presentation in a fourth portion, and

(d) the fourth portion configured to present, to the user, at least part of the individual document as both (1) selected in the third portion and (2) determined to be relevant to the individual reply;

responsive to selection of the individual textual conversation, providing the individual query as a prompt to the large language model;

obtaining the individual reply from the large language model, wherein the individual reply is presented in the second portion of the first user interface;

determining the individual document from the multiple documents that is relevant to the individual reply; and

presenting the individual document in the fourth portion of the first user interface, wherein a segment in the individual document that is relevant to the individual reply is emphasized in the fourth portion of the first user interface.

12 . The method of claim 11 , wherein the second portion is specific to the individual textual conversation from the set of textual conversations that has been selected through the first portion of the first user interface.

13 . The method of claim 11 , wherein the second portion presents modifications to the corpus of electronic documents.

14 . The method of claim 11 , wherein the second portion indicates whether the individual query was scoped to a subset of the corpus of electronic documents.

15 . The method of claim 11 , wherein the large language model is modeled such that the individual reply is limited in scope to one or more statements that have support in the corpus of electronic documents.

16 . The method of claim 11 , wherein the large language model is based on or derived from Generative Pre-trained Transformer 3 (GPT3).

17 . The method of claim 11 , wherein the second portion indicates a subset of the corpus of electronic documents determined to be relevant to the individual reply, wherein the subset includes the individual document.

18 . The method of claim 11 , wherein the individual reply is formatted for presentation based on the individual query.

19 . The method of claim 11 , wherein the segment in the individual document that is relevant to the individual reply is emphasized in the fourth portion of the first user interface by an indicator that indicates provenance of the individual reply.

20 . The method of claim 11 , wherein at least some parts of the first portion, the second portion, the third portion, and the fourth portion are presented to the user at the same time.