Same Entity, Different Script: Diagnosing Cross-Script Entity Matching Failures in Language Models
3 judges. Scored by all three judges. Only the day’s top stories get the claim check. How far stories are checked
This paper studies a weakness in multilingual language models: when the same person's name appears in Devanagari versus Latin script, models often fail to affirm the match.
Score
0.64total under rubric v2.0
- Significance0.57
- Reader interest0.65
- Audience breadth0.49
- Practical relevance0.60
- Substantiated0.88
- Technical depth0.95
- New information0.92
- Counts against: Promotional0.03
- Counts against: Overclaim0.26
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| Significance | 0.63 | 0.40 | 0.68 | 0.57 | 0.28 | +0.25 | +0.142 |
| Reader interest | 0.57 | 0.50 | 0.88 | 0.65 | 0.38 | +0.20 | +0.130 |
| Audience breadth | 0.54 | 0.35 | 0.58 | 0.49 | 0.23 | +0.10 | +0.049 |
| Practical relevance | 0.72 | 0.45 | 0.62 | 0.60 | 0.28 | +0.15 | +0.090 |
| Substantiated | 0.99 | 0.85 | 0.79 | 0.88 | 0.20 | +0.10 | +0.088 |
| Technical depth | 0.99 | 0.90 | 0.96 | 0.95 | 0.10 | +0.10 | +0.095 |
| New information | 0.90 | 0.90 | 0.97 | 0.92 | 0.07 | +0.10 | +0.092 |
| Promotional | 0.03 | 0.05 | 0.01 | 0.03 | 0.04 | −0.15 | −0.004 |
| Overclaim | 0.27 | 0.30 | 0.22 | 0.26 | 0.08 | −0.15 | −0.039 |
| Sum of contributions | +0.642 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.642 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN —
Recorded for context only, checked 24 Oct 2026, 08:54 UTC. It played no part in the ranking.