# Task-oriented Dialogue on Stanford Multi-Domain Dialogue (SMD) (test)

Source: [Task-oriented Dialogue on Stanford Multi-Domain Dialogue (SMD) (test)](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test>)

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

Data updated: 2026-02-26T05:36:50.528Z

Evaluation of Task-oriented Dialogue performance on the test split of the Stanford Multi-Domain Dialogue (SMD) dataset, measured using metrics like BLEU and domain-specific Entity F1 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 | Task-oriented Dialogue |
| Dataset (including any stated split) | Stanford Multi-Domain Dialogue (SMD) (test) |
| Metric | BLEU |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 31 of 31 |
| 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](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=BLEU>): Higher is better; unit: PERCENT.

- [Entity F1](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=Entity%20F1>): Higher is better; unit: PERCENT.

- [Scheduling Entity F1](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=Scheduling%20Entity%20F1>): Higher is better; unit: PERCENT.

- [Weather Entity F1](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=Weather%20Entity%20F1>): Higher is better; unit: PERCENT.

- [Navigation Entity F1](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=Navigation%20Entity%20F1>): Higher is better; unit: PERCENT.

- [Appropriateness Score](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=Appropriateness%20Score>): Higher is better; unit: SCALAR.

- [Humanlikeness Score](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=Humanlikeness%20Score>): Higher is better; unit: SCALAR.

- [Calendar Entity F1](<https://www.sota2.com/research/sota/task-oriented-dialogue-on-stanford-multi-domain-dialogue-smd-test.md?metric=Calendar%20Entity%20F1>): Higher is better; unit: PERCENT.

## Recorded results

| Method | BLEU | Result date | Report paper |
| --- | --- | --- | --- |
| COMET | 17.3 | Not recorded | 68fa6af21e209131cea2b447 |
| NS-Dial | 15.6 | Not recorded | 68fa6bfd1e209131cea317a1 |
| NS-Dial | 14.9 | Not recorded | 68fa6bfd1e209131cea317a1 |
| GLMP K3 | 14.79 | Not recorded | 68fa658c1e209131cea0b0ad |
| NS-Dial | 14.5 | Not recorded | 68fa6bfd1e209131cea317a1 |
| DF-Net | 14.4 | Not recorded | 68fa6af21e209131cea2b447 |
| DF-Net | 14.4 | Not recorded | 68fa6bfd1e209131cea317a1 |
| GraphDialog | 14.2 | Not recorded | 68fa6bfd1e209131cea317a1 |
| KB-Transformer | 13.9 | Not recorded | 68fa6af21e209131cea2b447 |
| KB-Retriever | 13.9 | Not recorded | 68fa6af21e209131cea2b447 |
| GLMP | 13.9 | Not recorded | 68fa6af21e209131cea2b447 |
| GLMP | 13.9 | Not recorded | 68fa6bfd1e209131cea317a1 |
| GLMP K1 | 13.83 | Not recorded | 68fa658c1e209131cea0b0ad |
| GraphDialog | 13.7 | Not recorded | 68fa6af21e209131cea2b447 |
| Human Performance | 13.5 | Not recorded | 68fa63841e209131ce9ff2fe |
| KV Retrieval Net | 13.2 | Not recorded | 68fa63841e209131ce9ff2fe |
| KVR | 13.2 | Not recorded | 68fa658c1e209131cea0b0ad |
| Mem2Seq | 12.6 | Not recorded | 68fa658c1e209131cea0b0ad |
| Mem2Seq | 12.6 | Not recorded | 68fa6af21e209131cea2b447 |
| Mem2Seq | 12.6 | Not recorded | 68fa6bfd1e209131cea317a1 |
| GLMP K6 | 12.37 | Not recorded | 68fa658c1e209131cea0b0ad |
| Copy Net | 11 | Not recorded | 68fa63841e209131ce9ff2fe |
| KV Retrieval Net | 10.8 | Not recorded | 68fa63841e209131ce9ff2fe |
| Attn. Seq2Seq | 10.2 | Not recorded | 68fa63841e209131ce9ff2fe |
| S2S + Attn | 9.3 | Not recorded | 68fa658c1e209131cea0b0ad |
| S2S | 8.4 | Not recorded | 68fa658c1e209131cea0b0ad |
| Ptr-Unk | 8.3 | Not recorded | 68fa658c1e209131cea0b0ad |
| Rule-Based | 6.6 | Not recorded | 68fa63841e209131ce9ff2fe |
| Rule-Based | 6.6 | Not recorded | 68fa658c1e209131cea0b0ad |

### COMET

BLEU: 17.3. Result ID: 698e44d1ffc3901de09c2d61.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### NS-Dial

BLEU: 15.6. Result ID: 698eb512ffc3901de09ff8e2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Implementation](<https://github.com/shiquanyang/NS-Dial>)

Recorded evaluation configuration: {"maximum depth of HRE module (D)":"3"}. This record does not assert independent reproduction.

### NS-Dial

BLEU: 14.9. Result ID: 698eb512ffc3901de09ff8e1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Implementation](<https://github.com/shiquanyang/NS-Dial>)

Recorded evaluation configuration: {"maximum depth of HRE module (D)":"1"}. This record does not assert independent reproduction.

### GLMP K3

BLEU: 14.79. Result ID: 698cb5b88b517eb4331c3450.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Implementation](<https://github.com/jasonwu0731/GLMP>)

Recorded evaluation configuration: {"K (Number of hops)":3}. This record does not assert independent reproduction.

### NS-Dial

BLEU: 14.5. Result ID: 698eb512ffc3901de09ff8e3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Implementation](<https://github.com/shiquanyang/NS-Dial>)

Recorded evaluation configuration: {"maximum depth of HRE module (D)":"5"}. This record does not assert independent reproduction.

### DF-Net

BLEU: 14.4. Result ID: 698e44d1ffc3901de09c2d5f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67b71e209131cea180fb.md>)

[Implementation](<https://github.com/LooperXX/DF-Net>)

Recorded evaluation configuration: {"Knowledge Representation":"Triplet"}. This record does not assert independent reproduction.

### DF-Net

BLEU: 14.4. Result ID: 698eb512ffc3901de09ff8e0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67b71e209131cea180fb.md>)

[Implementation](<https://github.com/LooperXX/DF-Net>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### GraphDialog

BLEU: 14.2. Result ID: 698eb512ffc3901de09ff8df.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68701e209131cea1c3d4.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### KB-Transformer

BLEU: 13.9. Result ID: 698e44d1ffc3901de09c2d5c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

Recorded evaluation configuration: {"Knowledge Representation":"Triplet"}. This record does not assert independent reproduction.

### KB-Retriever

BLEU: 13.9. Result ID: 698e44d1ffc3901de09c2d5d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66631e209131cea100ab.md>)

[Implementation](<https://github.com/yizhen20133868/Retriever-Dialogue>)

Recorded evaluation configuration: {"Knowledge Representation":"Row-entity"}. This record does not assert independent reproduction.

### GLMP

BLEU: 13.9. Result ID: 698e44d1ffc3901de09c2d5e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Implementation](<https://github.com/jasonwu0731/GLMP>)

Recorded evaluation configuration: {"Knowledge Representation":"Triplet"}. This record does not assert independent reproduction.

### GLMP

BLEU: 13.9. Result ID: 698eb512ffc3901de09ff8de.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Implementation](<https://github.com/jasonwu0731/GLMP>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### GLMP K1

BLEU: 13.83. Result ID: 698cb5b88b517eb4331c344f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Implementation](<https://github.com/jasonwu0731/GLMP>)

Recorded evaluation configuration: {"K (Number of hops)":1}. This record does not assert independent reproduction.

### GraphDialog

BLEU: 13.7. Result ID: 698e44d1ffc3901de09c2d60.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68701e209131cea1c3d4.md>)

Recorded evaluation configuration: {"Knowledge Representation":"Graph"}. This record does not assert independent reproduction.

### Human Performance

BLEU: 13.5. Result ID: 698bf7b5f729bc7f68e6f20f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### KV Retrieval Net

BLEU: 13.2. Result ID: 698bf7b5f729bc7f68e6f20e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Implementation](<https://nlp.stanford.edu/blog/a-new-multi-turn-multi-domain-task-oriented-dialogue-dataset/>)

Recorded evaluation configuration: {"encoder attention":true}. This record does not assert independent reproduction.

### KVR

BLEU: 13.2. Result ID: 698cb5b88b517eb4331c344a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Implementation](<https://nlp.stanford.edu/blog/a-new-multi-turn-multi-domain-task-oriented-dialogue-dataset/>)

Recorded evaluation configuration: {"Canonicalized forms":true}. This record does not assert independent reproduction.

### Mem2Seq

BLEU: 12.6. Result ID: 698cb5b88b517eb4331c344e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa645e1e209131cea040e1.md>)

[Implementation](<https://github.com/HLTCHKUST/Mem2Seq>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Mem2Seq

BLEU: 12.6. Result ID: 698e44d1ffc3901de09c2d5b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6af21e209131cea2b447.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa645e1e209131cea040e1.md>)

[Implementation](<https://github.com/HLTCHKUST/Mem2Seq>)

Recorded evaluation configuration: {"Knowledge Representation":"Triplet"}. This record does not assert independent reproduction.

### Mem2Seq

BLEU: 12.6. Result ID: 698eb512ffc3901de09ff8dd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6bfd1e209131cea317a1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa645e1e209131cea040e1.md>)

[Implementation](<https://github.com/HLTCHKUST/Mem2Seq>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### GLMP K6

BLEU: 12.37. Result ID: 698cb5b88b517eb4331c3451.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Implementation](<https://github.com/jasonwu0731/GLMP>)

Recorded evaluation configuration: {"K (Number of hops)":6}. This record does not assert independent reproduction.

### Copy Net

BLEU: 11. Result ID: 698bf7b5f729bc7f68e6f20b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63881e209131ce9ff57c.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### KV Retrieval Net

BLEU: 10.8. Result ID: 698bf7b5f729bc7f68e6f20d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Implementation](<https://nlp.stanford.edu/blog/a-new-multi-turn-multi-domain-task-oriented-dialogue-dataset/>)

Recorded evaluation configuration: {"encoder attention":false}. This record does not assert independent reproduction.

### Attn. Seq2Seq

BLEU: 10.2. Result ID: 698bf7b5f729bc7f68e6f20c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa627b1e209131ce9f9741.md>)

[Implementation](<https://github.com/lisa-groundhog/GroundHog>)

Recorded evaluation configuration: {"encoder attention":true}. This record does not assert independent reproduction.

### S2S + Attn

BLEU: 9.3. Result ID: 698cb5b88b517eb4331c344c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### S2S

BLEU: 8.4. Result ID: 698cb5b88b517eb4331c344b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Ptr-Unk

BLEU: 8.3. Result ID: 698cb5b88b517eb4331c344d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62b91e209131ce9fadbe.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Rule-Based

BLEU: 6.6. Result ID: 698bf7b5f729bc7f68e6f20a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Rule-Based

BLEU: 6.6. Result ID: 698cb5b88b517eb4331c3449.

[Reporting paper](<https://www.sota2.com/research/paper/68fa658c1e209131cea0b0ad.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63841e209131ce9ff2fe.md>)

[Implementation](<https://nlp.stanford.edu/blog/a-new-multi-turn-multi-domain-task-oriented-dialogue-dataset/>)

Recorded evaluation configuration: {"Canonicalized forms":true}. This record does not assert independent reproduction.

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