# Event Extraction on ACE (test)

Source: [Event Extraction on ACE (test)](<https://www.sota2.com/research/sota/event-extraction-on-ace-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:58:11.014Z

Evaluation of Event Extraction performance on the test split of the ACE corpus, measured using Practical and Full evaluation metrics including EM, COR, and SemSim scores.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Event Extraction |
| Dataset (including any stated split) | ACE (test) |
| Metric | Practical EM |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Practical EM](<https://www.sota2.com/research/sota/event-extraction-on-ace-test.md?metric=Practical%20EM>): Higher is better; unit: PERCENT.

- [Practical COR](<https://www.sota2.com/research/sota/event-extraction-on-ace-test.md?metric=Practical%20COR>): Higher is better; unit: PERCENT.

- [Practical SemSim](<https://www.sota2.com/research/sota/event-extraction-on-ace-test.md?metric=Practical%20SemSim>): Higher is better; unit: PERCENT.

- [Full EM](<https://www.sota2.com/research/sota/event-extraction-on-ace-test.md?metric=Full%20EM>): Higher is better; unit: PERCENT.

- [Full COR](<https://www.sota2.com/research/sota/event-extraction-on-ace-test.md?metric=Full%20COR>): Higher is better; unit: PERCENT.

- [Full SemSim](<https://www.sota2.com/research/sota/event-extraction-on-ace-test.md?metric=Full%20SemSim>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Practical EM | Result date | Report paper |
| --- | --- | --- | --- |
| RLQG | 41.39 | Not recorded | 68fa74361e209131cea60793 |
| SFT (Dynamic) | 39.31 | Not recorded | 68fa74361e209131cea60793 |
| GPT-4 (5shot) | 39.24 | Not recorded | 68fa74361e209131cea60793 |
| Dynamic-Q | 38.7 | Not recorded | 68fa74361e209131cea60793 |
| SFT (Guideline) | 38.62 | Not recorded | 68fa74361e209131cea60793 |
| Guideline-Q | 38.51 | Not recorded | 68fa74361e209131cea60793 |
| SFT (Back-Trans) | 38.24 | Not recorded | 68fa74361e209131cea60793 |
| SFT (Standard) | 37.63 | Not recorded | 68fa74361e209131cea60793 |
| Standard-Q | 37.42 | Not recorded | 68fa74361e209131cea60793 |
| Back-Trans-Q | 36.13 | Not recorded | 68fa74361e209131cea60793 |
| Simple-Q | 35.41 | Not recorded | 68fa74361e209131cea60793 |
| GPT-4 (0shot) | 28.97 | Not recorded | 68fa74361e209131cea60793 |
| LLaMA-2-13b-Chat (5shot) | 27.97 | Not recorded | 68fa74361e209131cea60793 |
| LLaMA-2-13b-Chat (0shot) | 1.21 | Not recorded | 68fa74361e209131cea60793 |

### RLQG

Practical EM: 41.39. Result ID: 699335de3a6c3148491475d4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

Recorded evaluation configuration: {"Method Category":"Reinforcement Learning Question Generation (Proposed)"}. This record does not assert independent reproduction.

### SFT (Dynamic)

Practical EM: 39.31. Result ID: 699335de3a6c3148491475cf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

Recorded evaluation configuration: {"Method Category":"Supervised Fine-tuning","Prompt Template Type":"Dynamic"}. This record does not assert independent reproduction.

### GPT-4 (5shot)

Practical EM: 39.24. Result ID: 699335de3a6c3148491475d3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa727a1e209131cea56a1e.md>)

[Implementation](<https://github.com/openai/evals>)

Recorded evaluation configuration: {"Method Category":"In-context learning","Backbone Model":"GPT-4","Number of Shots":"5"}. This record does not assert independent reproduction.

### Dynamic-Q

Practical EM: 38.7. Result ID: 699335de3a6c3148491475cb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70071e209131cea48b3f.md>)

[Implementation](<https://github.com/dataminr-ai/Event-Extraction-as-Question-Generation-and-Answering>)

Recorded evaluation configuration: {"Method Category":"Template","Prompt Template Type":"Dynamic"}. This record does not assert independent reproduction.

### SFT (Guideline)

Practical EM: 38.62. Result ID: 699335de3a6c3148491475ce.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

Recorded evaluation configuration: {"Method Category":"Supervised Fine-tuning","Prompt Template Type":"Guideline"}. This record does not assert independent reproduction.

### Guideline-Q

Practical EM: 38.51. Result ID: 699335de3a6c3148491475ca.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa693a1e209131cea21007.md>)

Recorded evaluation configuration: {"Method Category":"Template","Prompt Template Type":"Guideline"}. This record does not assert independent reproduction.

### SFT (Back-Trans)

Practical EM: 38.24. Result ID: 699335de3a6c3148491475cd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

Recorded evaluation configuration: {"Method Category":"Supervised Fine-tuning","Prompt Template Type":"Back-Translation"}. This record does not assert independent reproduction.

### SFT (Standard)

Practical EM: 37.63. Result ID: 699335de3a6c3148491475cc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

Recorded evaluation configuration: {"Method Category":"Supervised Fine-tuning","Prompt Template Type":"Standard"}. This record does not assert independent reproduction.

### Standard-Q

Practical EM: 37.42. Result ID: 699335de3a6c3148491475c8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa693a1e209131cea21007.md>)

Recorded evaluation configuration: {"Method Category":"Template","Prompt Template Type":"Standard"}. This record does not assert independent reproduction.

### Back-Trans-Q

Practical EM: 36.13. Result ID: 699335de3a6c3148491475c9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

Recorded evaluation configuration: {"Method Category":"Template","Prompt Template Type":"Back-Translation"}. This record does not assert independent reproduction.

### Simple-Q

Practical EM: 35.41. Result ID: 699335de3a6c3148491475c7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

Recorded evaluation configuration: {"Method Category":"Template","Prompt Template Type":"Simple"}. This record does not assert independent reproduction.

### GPT-4 (0shot)

Practical EM: 28.97. Result ID: 699335de3a6c3148491475d2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa727a1e209131cea56a1e.md>)

[Implementation](<https://github.com/openai/evals>)

Recorded evaluation configuration: {"Method Category":"In-context learning","Backbone Model":"GPT-4","Number of Shots":"0"}. This record does not assert independent reproduction.

### LLaMA-2-13b-Chat (5shot)

Practical EM: 27.97. Result ID: 699335de3a6c3148491475d1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701f1e209131cea492fe.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"Method Category":"In-context learning","Backbone Model":"LLaMA-2-13b-Chat","Number of Shots":"5"}. This record does not assert independent reproduction.

### LLaMA-2-13b-Chat (0shot)

Practical EM: 1.21. Result ID: 699335de3a6c3148491475d0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74361e209131cea60793.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701f1e209131cea492fe.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"Method Category":"In-context learning","Backbone Model":"LLaMA-2-13b-Chat","Number of Shots":"0"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
