<p>Scientific contributions are a direct reflection of a research paper's value, demonstrating its impact on existing theories or practices. Existing research mainly focuses on the authors' perceived or self-identified contributions, while the actual contributions which are the contributions of papers to other research in the context of scholarly communication are rarely investigated. This research studies the actual contributions of papers and further explores the labor input patterns of which from an input–output perspective, based on papers published in <i>Nature</i> and <i>Science</i> using 1.53 million citation contexts from citing literature. Additionally, we design a method for identifying the types of scientific contributions using large language model technology. Results show that the distribution of the actual contributions of studied papers is unbalanced, with experimental contributions being predominant, contrasting with majority of the theoretical and methodological contributions self-identified by authors, which highlights a notable discrepancy between actual contributions and authors' self-perceptions, indicating an "ideal bias." Regarding the input–output patterns of actual contributions, there is no significant correlation between the overall labor input pattern and the actual contribution pattern of papers, but a positive correlation is observed between input and output for specific types of scientific contributions, reflecting a "more effort, more gain" effect. As for the internal relationships among different types of scientific contributions, different types of DOL input in papers exhibit a notable co-occurrence trend; while, once the paper reaches the dissemination stage, the co-occurrence of different types of actual contributions becomes weaker, indicating that a paper’s actual contributions often concentrate on a single type.</p>