Username: EvalGPT Password: EvalGPTAga12345 After the login, you'll see a page where you enter details regarding the project (currently we support scientific research projects) that you are proposing. We support Turkish and English languages. All fields are required. At the bottom, there is a button called "Create Evaluation Report". It creates several LLM based judges behind the scenes. Each judge evaluates the project and write a report about the proposal. After that As the second phase, an LLM based supreme council is created. This council evaluates the reports of judges and make the final decision about whether to accept or reject the project. A final report is generated for the applicant with the rejection/acceptance details.
About
We are building an LLM based system that will evaluate the scientific project proposals based on several criteria( such as novelty, weaknesses, the problem it is trying to solve, method, project management, Impact & Value Added) and generate a detailed report whether/why it is suitable or not. The system will create several judges (or judge personas) that will act as members of a scientific committee. These LLM based judges will be created based on the keywords extracted from project proposal by using prompt engineering techniques. Each judge will have different field expertise, age, industry. As the second phase, an LLM based supreme council will be created. This council will evaluate the reports of judges and make the final decision about whether to accept or reject the project. A final report will be generated for the applicant with the rejection/acceptance details.The system will be flexible so that it can be used in any other fields where something (an idea, a software project, hackathon finalists, an illness, a debate and so on) should be evaluated by a committee. It can also be integrated to platforms such as backdropbuild and the ideas can be pre-screened automatically.
