Large-scale text analysis using generative language models: A case study in discovering public value expressions in AI patents
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract We put forward a novel approach using a generative language model (GPT-4) to produce labels and rationales for large-scale text analysis. The approach is used to discover public value expressions in patents. Using text (5.4 million sentences) for 154,934 US AI patent documents from the United States Patent and Trademark Office (USPTO), we design a semi-automated, human-supervised framework for identifying and labeling public value expressions in these sentences. A GPT-4 prompt is developed that includes definitions, guidelines, examples, and rationales for text classification. We evaluate the labels and rationales produced by GPT-4 using BLEU scores and topic modeling, finding that they are accurate, diverse, and faithful. GPT-4 achieved an advanced recognition of public value expressions from our framework, which it also uses to discover unseen public value expressions. The GPT-produced labels are used to train BERT-based classifiers and predict sentences on the entire database, achieving high F1 scores for the 3-class (0.85) and 2-class classification (0.91) tasks. We discuss the implications of our approach for conducting large-scale text analyses with complex and abstract concepts. With careful framework design and interactive human oversight, we suggest that generative language models can offer significant assistance in producing labels and rationales.
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.
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 | 0.005 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.011 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.002 |
| Open science | 0.000 | 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 it