Spatio-temporal distribution of global stromatolites through geological time identified by a large language model approach
Bibliographic record
Abstract
Introduction A substantial amount of data embedded within diverse literature makes it time-consuming to manually extract and compile extensive datasets. The use of large language models has become essential for the efficient extraction and analysis of big data. This study utilizes ChatGPT-4 to reconstruct a global database of stromatolites, spanning from the Precambrian to the present, to enhance our understanding of their spatial and temporal dynamics throughout geological time. Methods The data extraction process involved several steps: First, PDF documents containing stromatolite literature were gathered and converted into text format. Second, ChatGPT-4 was employed to extract data on stromatolite occurrences, including locations, ages, strata, and facies types from each sentence in the documents. Third, duplicates were removed, and the data were organized into three categories: 3,248 unique location-age pairs, 2,723 strata-age pairs, and 1,723 strata-age-facies type combinations. Additionally, 2,565 paleogeographical locations of stromatolite-bearing rocks were reconstructed using modern latitude and longitude coordinates and corresponding Phanerozoic ages. Results The newly obtained dataset reveals that stromatolite occurrences peaked during the Proterozoic, declined during the Early Phanerozoic, and exhibited fluctuations throughout the Phanerozoic. Seven global stromatolite hotspots were identified: the United States, Australia, India, Canada, China, England, and Russia. From the Cambrian to the Jurassic, stromatolites were predominantly distributed in low and middle latitudes, shifting to higher latitudes from the Cretaceous to the Quaternary. The proportion of inland aquatic stromatolites relative to marine stromatolites varied, ranging from 10% to 30% from the Mesoarchean to the Middle Mesoproterozoic, decreasing to less than 10% from the Late Mesoproterozoic to the Early Paleozoic, increasing to 10%–30% from the Devonian to the Jurassic, and remaining high (39%–53%) from the Cretaceous to the Quaternary. Discussion The findings highlight the temporal and spatial variability of stromatolite occurrences, shedding light on the evolution of these microbial structures over geological time. The distribution patterns suggest significant shifts in environmental conditions and provide valuable insights into paleogeographical and ecological dynamics. The use of ChatGPT-4 to extract and organize data from a large body of literature demonstrates the potential of large language models for advancing research in paleobiology and geology.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".