# Multi-task Reaction Prediction on Aggregated Multi-task Chemical Reaction Dataset (test)

Source: [Multi-task Reaction Prediction on Aggregated Multi-task Chemical Reaction Dataset (test)](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test>)

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

Data updated: 2026-02-26T05:37:08.505Z

Evaluation of models performing multi-task reaction prediction on the test split of the Aggregated Multi-task Chemical Reaction Dataset, using sequence generation metrics (BLEU, METEOR) and chemical validity metrics.

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 | Multi-task Reaction Prediction |
| Dataset (including any stated split) | Aggregated Multi-task Chemical Reaction Dataset (test) |
| Metric | BLEU Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [BLEU Score](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test.md?metric=BLEU%20Score>): Higher is better; unit: SCALAR.

- [METEOR Score](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test.md?metric=METEOR%20Score>): Higher is better; unit: SCALAR.

- [Exact Match Score](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test.md?metric=Exact%20Match%20Score>): Higher is better; unit: SCALAR.

- [Similarity Score](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test.md?metric=Similarity%20Score>): Higher is better; unit: SCALAR.

- [Validity Score](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test.md?metric=Validity%20Score>): Higher is better; unit: PERCENT.

- [Improvement Metric](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test.md?metric=Improvement%20Metric>): Higher is better; unit: SCALAR.

## Recorded results

| Method | BLEU Score | Result date | Report paper |
| --- | --- | --- | --- |
| LLM-CRP (RT+IA, N=10) | 0.879 | Not recorded | 68fa5f581e209131ce9e9189 |
| MolT5 | 0.837 | Not recorded | 68fa5f581e209131ce9e9189 |
| T5 | 0.825 | Not recorded | 68fa5f581e209131ce9e9189 |
| Text+Chem T5 | 0.822 | Not recorded | 68fa5f581e209131ce9e9189 |
| Text+Chem T5 | 0.799 | Not recorded | 68fa5f581e209131ce9e9189 |

### LLM-CRP (RT+IA, N=10)

BLEU Score: 0.879. Result ID: 698bc4097a967eddb77131c5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5f581e209131ce9e9189.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa5f581e209131ce9e9189.md>)

[Implementation](<https://github.com/AI-HPC-Research-Team/SLM4CRP>)

Recorded evaluation configuration: {"Reaction Type integration":true,"Instruction Adaptation":true,"Cluster Number":10}. This record does not assert independent reproduction.

### MolT5

BLEU Score: 0.837. Result ID: 698bc4097a967eddb77131c2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5f581e209131ce9e9189.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6dde1e209131cea3c90f.md>)

[Implementation](<https://github.com/blender-nlp/MolT5>)

Recorded evaluation configuration: {"Model":"MolT5"}. This record does not assert independent reproduction.

### T5

BLEU Score: 0.825. Result ID: 698bc4097a967eddb77131c1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5f581e209131ce9e9189.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa709e1e209131cea4c470.md>)

[Implementation](<https://github.com/google-research/text-to-text-transfer-transformer>)

Recorded evaluation configuration: {"Model":"T5"}. This record does not assert independent reproduction.

### Text+Chem T5

BLEU Score: 0.822. Result ID: 698bc4097a967eddb77131c3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5f581e209131ce9e9189.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f711e209131cea459c6.md>)

[Implementation](<https://github.com/GT4SD/multitask_text_and_chemistry_t5>)

Recorded evaluation configuration: {"Training Mode":"finetune"}. This record does not assert independent reproduction.

### Text+Chem T5

BLEU Score: 0.799. Result ID: 698bc4097a967eddb77131c4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5f581e209131ce9e9189.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f711e209131cea459c6.md>)

[Implementation](<https://github.com/GT4SD/multitask_text_and_chemistry_t5>)

Recorded evaluation configuration: {"Evaluation Protocol":"avg-tasks (baseline)"}. This record does not assert independent reproduction.

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