# Abstract Retrieval on Synthetic Tasks (EM)

Source: [Abstract Retrieval on Synthetic Tasks (EM)](<https://www.sota2.com/research/sota/abstract-retrieval-on-synthetic-tasks-em>)

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

Data updated: 2026-02-26T05:42:27.482Z

Evaluation of Abstract Retrieval performance on the Synthetic Tasks dataset, measured using Exact Match (EM).

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 | Abstract Retrieval |
| Dataset (including any stated split) | Synthetic Tasks |
| Metric | Exact Match |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [Exact Match](<https://www.sota2.com/research/sota/abstract-retrieval-on-synthetic-tasks-em.md?metric=Exact%20Match>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Exact Match | Result date | Report paper |
| --- | --- | --- | --- |
| Our model | 98.5 | Not recorded | 68fa746e1e209131cea61af7 |
| ChatGLM3-6B-32K | 94 | Not recorded | 68fa746e1e209131cea61af7 |
| Baichuan2-Turbo-192K | 90 | Not recorded | 68fa746e1e209131cea61af7 |
| GPT3.5-Turbo-16K | 77.5 | Not recorded | 68fa746e1e209131cea61af7 |
| ChatGLM2-6B-32K | 64.5 | Not recorded | 68fa746e1e209131cea61af7 |
| Qwen-14B-Chat | 40 | Not recorded | 68fa746e1e209131cea61af7 |
| Longchat-v1.5-7B-32K | 7.6 | Not recorded | 68fa746e1e209131cea61af7 |
| Vicuna-v1.5-7B-16K | 5 | Not recorded | 68fa746e1e209131cea61af7 |
| Xgen-7B-8K | 3.5 | Not recorded | 68fa746e1e209131cea61af7 |
| InternLM-7B-8K | 0.9 | Not recorded | 68fa746e1e209131cea61af7 |

### Our model

Exact Match: 98.5. Result ID: 699347ca3a6c31484914e362.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

[Implementation](<https://huggingface.co/IDEA-CCNL/Ziya-Reader-13B-v1.0>)

Recorded evaluation configuration: {"Context Window Size":"8K"}. This record does not assert independent reproduction.

### ChatGLM3-6B-32K

Exact Match: 94. Result ID: 699347ca3a6c31484914e35c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c111e209131cea31e21.md>)

[Implementation](<https://github.com/THUDM/GLM>)

Recorded evaluation configuration: {"Context Window Size":"32K","Backbone":"6B"}. This record does not assert independent reproduction.

### Baichuan2-Turbo-192K

Exact Match: 90. Result ID: 699347ca3a6c31484914e359.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70aa1e209131cea4c547.md>)

Recorded evaluation configuration: {"Context Window Size":"192K"}. This record does not assert independent reproduction.

### GPT3.5-Turbo-16K

Exact Match: 77.5. Result ID: 699347ca3a6c31484914e35d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

Recorded evaluation configuration: {"Context Window Size":"16K"}. This record does not assert independent reproduction.

### ChatGLM2-6B-32K

Exact Match: 64.5. Result ID: 699347ca3a6c31484914e35b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c111e209131cea31e21.md>)

[Implementation](<https://github.com/THUDM/GLM>)

Recorded evaluation configuration: {"Context Window Size":"32K","Backbone":"6B"}. This record does not assert independent reproduction.

### Qwen-14B-Chat

Exact Match: 40. Result ID: 699347ca3a6c31484914e361.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

Recorded evaluation configuration: {"Context Window Size":"8K","Backbone":"14B"}. This record does not assert independent reproduction.

### Longchat-v1.5-7B-32K

Exact Match: 7.6. Result ID: 699347ca3a6c31484914e35a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

Recorded evaluation configuration: {"Context Window Size":"32K","Backbone":"7B"}. This record does not assert independent reproduction.

### Vicuna-v1.5-7B-16K

Exact Match: 5. Result ID: 699347ca3a6c31484914e35e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71aa1e209131cea526df.md>)

[Implementation](<https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge>)

Recorded evaluation configuration: {"Context Window Size":"16K","Backbone":"7B"}. This record does not assert independent reproduction.

### Xgen-7B-8K

Exact Match: 3.5. Result ID: 699347ca3a6c31484914e35f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

Recorded evaluation configuration: {"Context Window Size":"8K","Backbone":"7B"}. This record does not assert independent reproduction.

### InternLM-7B-8K

Exact Match: 0.9. Result ID: 699347ca3a6c31484914e360.

[Reporting paper](<https://www.sota2.com/research/paper/68fa746e1e209131cea61af7.md>)

Recorded evaluation configuration: {"Context Window Size":"8K","Backbone":"7B"}. This record does not assert independent reproduction.

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