Back to results
US 2026/0244831 A1Published application
Argument optimization and feedback detection for llm simulation evaluation
AssigneeVerdicthub
InventorsSWARUP ACHARYA, GAURAV TIWARI, TROY HIRSCHHORN
Life of this patent
Estimating the term…
Abstract
LLM improvements are disclosed relating to simulated dispute evaluation. By establishing a stateless digital dispute evaluator, multiple different scenarios can be evaluated. Such scenarios may differ as to their specifics (e.g. arguments applicable in one scenario vs. another scenario). Differing argument parameters may lead to disparate simulated evaluation scenario outcomes. Outcome optimization information may be generated as a result of simulated evaluation analysis where multiple scenarios may be present.
Claims20
A system, comprising:
- a processor;
- and
- a non-transitory computer-readable medium having stored thereon instructions that are executable by the processor to cause the system to perform operations comprising:
- determining a first set of legal arguments applicable to a first party in regard to a legal dispute involving the first party;
- performing a first simulation, using a first large language model (LLM), of a dispute evaluation process based on the first set of legal arguments, wherein the first simulation uses a dispute evaluator comprising one or more digital entities each having a respective set of data characteristics;
- determining a second set of legal arguments applicable to the first party in regard to the legal dispute, wherein the second set of legal arguments differs from the first set of legal arguments with respect to at least one argument parameter;
- performing a second simulation, using a second LLM, of the dispute evaluation process using the dispute evaluator comprising the one or more digital entities;
- determining a difference between an outcome of the first simulation and the second simulation;
- and
- generating argument optimization outcome information based on the difference.
7 more elements
A method, comprising:
- determining, by a computer system, a first set of legal arguments applicable to a first party in regard to a legal dispute involving the first party;
- performing a first simulation, using a first large language model (LLM), of a dispute evaluation process based on the first set of legal arguments, wherein the first simulation uses a dispute evaluator comprising one or more digital entities each having a respective set of data characteristics;
- determining, by the computer system, a second set of legal arguments applicable to the first party in regard to the legal dispute, wherein the second set of legal arguments differs from the first set of legal arguments with respect to at least one argument parameter;
- performing a second simulation, using a second LLM, of the dispute evaluation process using the dispute evaluator comprising the one or more digital entities;
- determining a difference between an outcome of the first simulation and the second simulation;
- and
- generating argument optimization outcome information based on the difference.
4 more elements
A non-transitory computer-readable medium having stored thereon instructions that are executable by a processor of a computer system to cause the computer system to perform operations comprising:
- determining a first set of legal arguments applicable to a first party in regard to a legal dispute involving the first party;
- performing a first simulation, using a first large language model (LLM), of a dispute evaluation process based on the first set of legal arguments, wherein the first simulation uses a dispute evaluator comprising one or more digital entities each having a respective set of data characteristics;
- determining a second set of legal arguments applicable to the first party in regard to the legal dispute, wherein the second set of legal arguments differs from the first set of legal arguments with respect to at least one argument parameter;
- performing a second simulation, using a second LLM, of the dispute evaluation process using the dispute evaluator comprising the one or more digital entities, wherein performing the second simulation comprises the dispute evaluator performing a current evaluation of the second set of legal arguments without the dispute evaluator retaining information regarding a previous evaluation of the first set of legal arguments, and wherein a composition of the dispute evaluator is not changed between performing the previous evaluation and the current evaluation;
- determining a difference between an outcome of the first simulation and the second simulation;
- and
- generating argument optimization outcome information based on the difference.
4 more elements
Dependencies are inferred from the claim text.