Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
csv
Languages:
English
Size:
< 1K
Tags:
clarus
clinical-trials
narrative-integrity
section-transition
evidence-reporting
clinical-audit
License:
sample_id string | source_framework string | target_framework string | concept string | source_definition string | target_definition string | claim string | option_a string | option_b string | task string | correct_option string | required_note string | required_keywords string | axis string | difficulty int64 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
CEFT-0001 | Protocol | Publication | Primary endpoint | PFS per RECIST 1.1 | OS at 12 months | Report OS as primary efficacy | Invalid correspondence, endpoint changed | Valid correspondence, same efficacy goal | Judge correspondence and write note | A | State PFS vs OS shift and flag endpoint substitution | PFS|OS|endpoint shift|protocol|publication | endpoint_identity | 3 |
CEFT-0002 | Protocol | Publication | Primary endpoint | Time to progression | PFS | Treat as same endpoint | Partial correspondence, definitions differ | Valid correspondence, equivalent metrics | Judge correspondence and write note | A | State TTP vs PFS difference and define event rules | TTP|PFS|event rule|difference|define | endpoint_definition | 2 |
CEFT-0003 | Protocol | SAP | Primary endpoint | ORR confirmed | ORR unconfirmed | Assume confirmation unchanged | Invalid correspondence, responder rule changed | Valid correspondence, same ORR | Judge correspondence and write note | A | State confirmed vs unconfirmed ORR and flag drift | confirmed|unconfirmed|ORR|responder|drift | responder_rule | 2 |
CEFT-0004 | SAP | Publication | Primary endpoint | PFS from randomization | PFS from first dose | Treat as same time origin | Partial correspondence, time origin differs | Valid correspondence, PFS label matches | Judge correspondence and write note | A | State randomization vs first dose and specify time zero | randomization|first dose|time zero|PFS|specify | time_origin | 2 |
CEFT-0005 | Protocol | Publication | Secondary endpoint | QoL mean change | QoL responder rate | Report as same patient benefit | Invalid correspondence, scale changed | Partial correspondence, map with transform | Judge correspondence and write note | A | State mean change vs responder rate and flag scale shift | mean change|responder rate|QoL|scale|shift | endpoint_scale | 3 |
CEFT-0006 | Protocol | Publication | Composite endpoint | MACE includes hospitalization | MACE excludes hospitalization | Treat composite as unchanged | Invalid correspondence, component removed | Partial correspondence, restate components | Judge correspondence and write note | A | State hospitalization component change and flag drift | MACE|hospitalization|component|change|drift | composite_components | 3 |
CEFT-0007 | Protocol | SAP | Endpoint hierarchy | Primary OS, secondary PFS | Co-primary OS and PFS | Assume hierarchy unchanged | Invalid correspondence, hierarchy altered | Valid correspondence, justified change | Judge correspondence and write note | A | State hierarchy change and require justification | primary|secondary|co-primary|hierarchy|justify | endpoint_hierarchy | 3 |
CEFT-0008 | Protocol | Publication | Endpoint window | 12-month OS | 24-month OS | Pool survival outcomes | Partial correspondence, window differs | Valid correspondence, same endpoint | Judge correspondence and write note | A | State 12 vs 24 month window and flag mismatch | 12 month|24 month|window|OS|mismatch | window_definition | 2 |
CEFT-0009 | Protocol | Publication | Exploratory endpoint | Biomarker response exploratory | Biomarker response efficacy claim | Treat exploratory as confirmatory | Invalid correspondence, role escalation | Valid correspondence, exploratory retained | Judge correspondence and write note | A | State exploratory to efficacy shift and flag escalation | exploratory|efficacy|biomarker|escalation|drift | endpoint_role | 3 |
CEFT-0010 | Protocol | Publication | Non-inferiority margin | NI margin HR 1.25 | NI margin HR 1.10 | Report same NI conclusion | Invalid correspondence, margin changed | Partial correspondence, restate margin | Judge correspondence and write note | A | State margin change and flag decision shift | NI margin|1.25|1.10|decision|shift | decision_threshold | 3 |
Clarus Clinical Narrative Integrity v0.1.1
What this dataset tests
- You track meaning as it moves across sections of a clinical paper
- You detect when claims shift strength or scope without justification
- You flag narrative drift introduced by section transitions
Scope
- One document
- Multiple internal sections
- Abstract
- Methods
- Results
- Discussion
- Conclusion
- One publication lifecycle
How this differs from v0.1
- v0.1 tests internal consistency within a single narrative fragment
- v0.1.1 tests continuity across narrative sections
- v0.1.1 is stricter and higher resolution
- Same invariant
- Finer audit surface
Typical failure modes
- Abstract upgrades exploratory findings
- Discussion removes qualifiers stated in Results
- Conclusion reframes safety or efficacy language
- Claims migrate from descriptive to causal
- Negative findings fade across sections
What this dataset does not test
- Protocol to publication transitions
- Registry or SAP drift
- Phase-to-phase definitional changes
Task format
- Each row presents paired or sequential narrative excerpts
- The model judges whether meaning is preserved across sections
- Output is a binary choice
Label meaning
- A indicates narrative integrity failure across sections
- B indicates narrative integrity preserved
Use cases
- Automated audit of clinical manuscripts
- Journal and peer-review triage
- Detection of rhetorical inflation in reporting
Version note
- v0.1 remains valid for single-fragment integrity checks
- v0.1.1 extends coverage to section-to-section narrative transitions
- Neither version supersedes the other
- Downloads last month
- 17