# Context-aware Instruction Following on Academic Paper Abstracts (test)

Source: [Context-aware Instruction Following on Academic Paper Abstracts (test)](<https://www.sota2.com/research/sota/context-aware-instruction-following-on-academic-paper-abstracts-test>)

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

Data updated: 2026-02-26T05:54:39.217Z

Evaluation of model performance on context-aware instruction following tasks using the test split of the Academic Paper Abstracts dataset, measured by Overlap metrics and Perplexity.

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 | Context-aware Instruction Following |
| Dataset (including any stated split) | Academic Paper Abstracts (test) |
| Metric | Overlap (w) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [Overlap (w)](<https://www.sota2.com/research/sota/context-aware-instruction-following-on-academic-paper-abstracts-test.md?metric=Overlap%20(w)>): Higher is better; unit: SCALAR.

- [Perplexity](<https://www.sota2.com/research/sota/context-aware-instruction-following-on-academic-paper-abstracts-test.md?metric=Perplexity>): Lower is better; unit: SCALAR.

- [Overlap (w/o)](<https://www.sota2.com/research/sota/context-aware-instruction-following-on-academic-paper-abstracts-test.md?metric=Overlap%20(w%2Fo)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Overlap (w) | Result date | Report paper |
| --- | --- | --- | --- |
| GPT-4-turbo | 8.64 | Not recorded | 68fa738b1e209131cea5d1d0 |
| Qwen-Chat(v1.5) + Qwen-Chat(v1.5) (Logit-based CoGenesis) | 8.04 | Not recorded | 68fa738b1e209131cea5d1d0 |

### GPT-4-turbo

Overlap (w): 8.64. Result ID: 6992ef833a6c31484912782d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

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

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

Recorded evaluation configuration: {"Mode":"zero-shot LLM with context (upper bound)"}. This record does not assert independent reproduction.

### Qwen-Chat(v1.5) + Qwen-Chat(v1.5) (Logit-based CoGenesis)

Overlap (w): 8.04. Result ID: 6992ef833a6c31484912782e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

Recorded evaluation configuration: {"Mode":"mixed-scale collaboration","Params":"mixed"}. This record does not assert independent reproduction.

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