The final LT4CPR Shared Task rankings are reported below. Complete scores are available in the downloadable workbook.
Official Shared Task Leaderboard
This page reports the official LT4CPR Shared Task rankings. The Primary System Ranking uses each team's designated primary submission, while the Best System Ranking uses the highest-scoring system submitted by each team.
← Back to the Shared Task page
Primary System Ranking
Each team is ranked using the system designated as primary in its submission.
| Rank | Team | Primary System | BERTScore F1 | BLEURT | Official Primary Score |
|---|---|---|---|---|---|
| 🥇 1 | Stockmark | report-refinement | 0.7269 | 0.6362 | 0.6816 |
| 🥈 2 | UTS | v6 | 0.7009 | 0.6275 | 0.6642 |
| 🥉 3 | Dy | rules-v2 | 0.6739 | 0.6265 | 0.6502 |
| 4 | Meng-qiu in Zhongguancun | grounded | 0.5839 | 0.6071 | 0.5955 |
| 5 | Amrood | nemotron-120b-sitrep | 0.5984 | 0.5483 | 0.5733 |
| 6 | gogogo | r01g-sr | 0.5018 | 0.5528 | 0.5273 |
| 7 | BMU | system_1 | 0.4511 | 0.5170 | 0.4841 |
| 8 | Neural Coder | extraction-de | 0.5082 | 0.4514 | 0.4798 |
| 9 | Databaes | g5-evidence-grounded | 0.3947 | 0.5153 | 0.4550 |
Best System Ranking
Each team is ranked using its highest-scoring submitted system, whether or not that system was designated primary.
| Rank | Team | Best System | Primary? | BERTScore F1 | BLEURT | Official Primary Score |
|---|---|---|---|---|---|---|
| 🥇 1 | Stockmark | report-refinement | Yes | 0.7269 | 0.6362 | 0.6816 |
| 🥈 2 | UTS | v6mbr | No | 0.7114 | 0.6221 | 0.6667 |
| 🥉 3 | Dy | rules-v2 | Yes | 0.6739 | 0.6265 | 0.6502 |
| 4 | Meng-qiu in Zhongguancun | cue | No | 0.5841 | 0.6071 | 0.5956 |
| 5 | Amrood | nemotron-120b-sitrep | Yes | 0.5984 | 0.5483 | 0.5733 |
| 6 | gogogo | r01g-sr | Yes | 0.5018 | 0.5528 | 0.5273 |
| 7 | BMU | system_1 | Yes | 0.4511 | 0.5170 | 0.4841 |
| 8 | Neural Coder | extraction-de | Yes | 0.5082 | 0.4514 | 0.4798 |
| 9 | Databaes | g5-evidence-grounded | Yes | 0.3947 | 0.5153 | 0.4550 |
Scores are rounded to four decimal places for display. Rankings use the unrounded official scores. The organizer baseline is available in the downloadable workbook but is not included in participant rankings.