Zwei Benchmarks: Schreibqualität und Fehlererkennung
Über 7 Fachgebiete, 130 verschiedene Publikationen und 280 feste Schreibpositionen erzielte Autocomplete mit Evidence einen fachübergreifenden Nettovorsprung von 31 Siegen; Evidence Card erkannte 91,4 % der sechs eindeutigen Fehlerarten.
AUTOCOMPLETE · THE SAME 280 FIRST ATTEMPTS
The with-card version was judged better in 138/280 first attempts
The purple bars below show the share of all 280 attempts
All frozen positions280/280 · 100%
Text plus model-reported card use140 of 280 tests produced text and reported card use. The previous 150 included 10 cases without continuation text; corrected on 15 October 2026.
50%140/280
Both wrote text; comparable pairs
107/280· 38%
That is 76% of the 140 adopted cases; in the other 33 (24%), the no-card arm was blank.
Use is reported by the generator, not an evidence-opportunity label.
WHEN CARDS WERE USED AND BOTH WROTE
With-card version preferred140 continuations with text reported card use; both sides produced text in 107. The no-card arm produced no text in the other 33. The comparison hides which version received cards so the label does not influence the judgment.
65%70/107
About the same7% · 7/107
Without cards better28% · 30/107
COMPARISON · NOT COUNTED AS VALID ADOPTION
39%55/140 with-card wins in the overall review
130 had no use report and 10 reported use without valid text; also 25 ties, 54 no-card wins, 3 neither usable, and 3 inconclusive. This group does not mean evidence was unsuitable.
RESULTS ACROSS SEVEN FIELDS
Card use was reported 140 times; 107 produced comparable text
The evaluation covers forty writing positions in each of seven fields. Card use is self-reported by the generator; the preferences below cover only tests where both sides produced text.
Field and sample
Model-reported card use140 of 280 tests produced text and reported card use. The previous 150 included 10 cases without continuation text; corrected on 15 October 2026.
Comparison when both versions produced text140 continuations with text reported card use; both sides produced text in 107. The no-card arm produced no text in the other 33. The comparison hides which version received cards so the label does not influence the judgment.
Model-reported card useThe with-card arm produced text and reported card use in 22 of 40 tests. Use is model-reported, not independently verified.
55%· 22/40
22 of 40 produced text and reported card use
When both versions wrote text, which was better?22 continuations with text reported card use; both sides produced text in 17. The no-card arm had no text in the other 5.
Model-reported card useThe with-card arm produced text and reported card use in 23 of 40 tests. Use is model-reported, not independently verified.
57%· 23/40
23 of 40 produced text and reported card use
The previous 25/40 included two with-card outputs without text; corrected here.
When both versions wrote text, which was better?23 continuations with text reported card use; both sides produced text in 16. The no-card arm had no text in the other 7.
Model-reported card useThe with-card arm produced text and reported card use in 23 of 40 tests. Use is model-reported, not independently verified.
57%· 23/40
23 of 40 produced text and reported card use
When both versions wrote text, which was better?23 continuations with text reported card use; both sides produced text in 13. The no-card arm had no text in the other 10.
Model-reported card useThe with-card arm produced text and reported card use in 25 of 40 tests. Use is model-reported, not independently verified.
63%· 25/40
25 of 40 produced text and reported card use
When both versions wrote text, which was better?25 continuations with text reported card use; both sides produced text in 19. The no-card arm had no text in the other 6.
Model-reported card useThe with-card arm produced text and reported card use in 14 of 40 tests. Use is model-reported, not independently verified.
35%· 14/40
14 of 40 produced text and reported card use
When both versions wrote text, which was better?14 continuations with text reported card use; both sides produced text in 9. The no-card arm had no text in the other 5.
Model-reported card useThe with-card arm produced text and reported card use in 18 of 40 tests. Use is model-reported, not independently verified.
45%· 18/40
18 of 40 produced text and reported card use
When both versions wrote text, which was better?18 continuations with text reported card use; both sides produced text in 18. The no-card arm had no text in the other 0.
Model-reported card useThe with-card arm produced text and reported card use in 15 of 40 tests. Use is model-reported, not independently verified.
38%· 15/40
15 of 40 produced text and reported card use
When both versions wrote text, which was better?15 continuations with text reported card use; both sides produced text in 15. The no-card arm had no text in the other 0.
Of 15 reported uses, 15 had text on both sides: 9 with-card wins, 4 ties, 2 no-card wins; the no-card arm had no text in 0.
What to know before reading the results
Positions fixed in advance: Forty positions were selected in each of seven fields; positions and cards were fixed before any outputs were generated.
Matched comparison: The writing position, model and task stay the same; only whether literature cards are supplied changes.
No repeated retries: Each side keeps its first output and is not regenerated based on the result.
Who judges quality: An AI judge compares anonymized outputs; this is not an independent human review.
Who reports card use: Card use is self-reported by the generation model without case-by-case human review.
Results are inspectable: All 280 cases show the draft, literature cards and first outcomes from both sides; judged cases also show the review rationale.
CASES FROM THIS EVALUATION
Explore all 280 autocomplete tests
Select any number above to filter to its cases. Each case shows the manuscript, cards and both first outcomes; judged cases also show the review rationale.
Preferences come from an anonymized AI judge without independent human review; four unresolved disagreements are listed as inconclusive.