Enhancing Legal Text Entailment: Evaluating Model Architectures, Training Approaches, and Interpretability
Bibliographic record
Abstract
This study investigates the effectiveness of various model architectures and training strategies for legal text classification, with a focus on entailment classification for case decisions from the Federal Court of Canada. We compare the performance of RoBERTa models with and without domain-specific further pretraining, to larger language models such as Llama 2, Llama 3, and GPT-4o adapted to the task by using prompt engineering and LORA fine-tuning. Additionally, we investigate different methods that can be used to explain the decisions of the models and evaluate their adequacy, understandability, trustworthiness, and sufficiency. Our findings suggest that for legal entailment classification, domain-specific pretraining can improve performance for smaller models, while larger language models show promise in outperforming prompt engineering for classification when fine-tuned with LoRA, as well as in generating more interpretable explanations. To the best of our knowledge, this is the first study in the context of Canadian legal AI to explore the effects of further pretraining on small and large language models, and the integration of language model adaptation and explainability into one system for legal text entailment classification.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.033 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.004 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".