[2505.03332] AI-Driven Scholarly Peer Review via Persistent Workflow Prompting, Meta-Prompting, and Meta-Reasoning

View a PDF file from the paper entitled Review of Scientific Poets that are driven by artificial intelligence through the continuous flow of workflow, connection of metal, expanded, by Evgeny Markhasin

PDF view

a summary:The critical peer review of scientific manuscripts is a major challenge to the LLMS models, partly due to data restrictions and the complexity of expert thinking. This report submits a PWP’s continuous workflow, which is a widely directed engineering systematically designed to fill this gap using Standard LLMs (zero code, without applications programming interfaces). We offer the PWP router proof of the concept of cash analysis of experimental chemistry manuscripts, and is characterized by a pyramid and sensual structure (organized by reducing mobility) that determines the functioning of detailed analysis. We develop this PWP router through the repetitive application of descriptive and expanded connection techniques that aim to record the work of experts review systematically, including implicit knowledge. It was presented once at the beginning of the session, this PWP router increases LLM with the ongoing workflow resulting from subsequent queries, and LLMS directing modern logic through multimedia systematic assessments. The LLM demonstrations of PWP that determine the main methodological defects appear in the case of a test with a reduction in the bias of LLM and the implementation of complex tasks, including distinguished claims of evidence, integration of text/image/form analysis of parameters conclusion, implementation of quantitative feasibility examination, comparing comparative estimates with claims, and evaluating pre -interview. To ensure transparency and facilitate symmetric copies, we make complete claims, detailed explanatory analyzes, and interactive chat records as complementary resources. In addition to the specific application, this work provides visions about the descriptive development process itself, with highlighting the PWP capabilities, informing the formal nature of the detailed workflow, to enable the advanced analysis using the LLMS easily available for complex scientific tasks.