Why T5 Forgets Entities: Diagnosing and Mitigating Attention Failures in Summarization
Bibliographic record
Abstract
While T5 achieves strong performance on text to-text tasks, its summarization behavior remains understud-ied-particularly in how attention mechanisms contribute to entity loss and hallucinations. We present the first layer-wise diagnosis of T5-base’s summarization pipeline, revealing that: (1) middle layers (5-8), critical for entity linking, systematically drop mid-text entities due to positional bias; (2) decoder biases disproportionately favor temporal markers (e.g., “Today”) over factual content. Using the CNN/DailyMail [1] dataset ($\mathbf{n} \boldsymbol{=} \mathbf{1 0 0}$), we quantify and evaluate these issues with novel metrics like Positional Entity Retention Score (PERS) and Hallucination Severity Index (HSI)-are introduced to diagnose entity loss and hallucination trends, offering innovative tools for evaluating summarization reliability. These findings inform practical strategies to enhance summarization pipelines
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 | 0.003 | 0.018 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 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 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.003 |
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".