# Dialog State Tracking on SpokenWoz (test) Source: [Dialog State Tracking on SpokenWoz (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:47:36.245Z Evaluation of Dialog State Tracking performance on the test split of the SpokenWoz dataset, typically measured using Joint Goal Accuracy (JGA). 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 | Dialog State Tracking | | Dataset (including any stated split) | SpokenWoz (test) | | Metric | JGA | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 28 of 28 | | 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 - [JGA](): Higher is better; unit: PERCENT. ## Recorded results | Method | JGA | Result date | Report paper | | --- | --- | --- | --- | | Full Spoken Context + Gemma2-9B-Instruct | 45.52 | Not recorded | 68fa7c0d1e209131cea8c22b | | Compressed Spoken Context + Gemma2-9B-Instruct | 43.16 | Not recorded | 68fa7c0d1e209131cea8c22b | | Joint Speech and Text E2E DST | 43 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 42.6 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 42.2 | Not recorded | 692d2147d6b0dd334293190a | | WavLM + conn. + Gemma2-9B-Instruct | 42.17 | Not recorded | 68fa7c0d1e209131cea8c22b | | Joint Speech and Text E2E DST | 41 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 40.3 | Not recorded | 692d2147d6b0dd334293190a | | Full Spoken Context | 39.32 | Not recorded | 68fa7c0d1e209131cea8c22b | | Joint Speech and Text E2E DST | 37.4 | Not recorded | 692d2147d6b0dd334293190a | | Compressed Spoken Context | 36.49 | Not recorded | 68fa7c0d1e209131cea8c22b | | Joint Speech and Text E2E DST | 36.3 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 35.6 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 34.9 | Not recorded | 692d2147d6b0dd334293190a | | WavLM + conn. + OLMo-1B | 34.66 | Not recorded | 68fa7c0d1e209131cea8c22b | | Joint Speech and Text E2E DST | 34.1 | Not recorded | 692d2147d6b0dd334293190a | | OLMo-1B \[7\] | 32.1 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 30.9 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 29.3 | Not recorded | 692d2147d6b0dd334293190a | | UBAR + GenWOZ | 25.9 | Not recorded | 68fa7c0d1e209131cea8c22b | | SPACE+WavLMalign | 25.65 | Not recorded | 68fa7c0d1e209131cea8c22b | | E2E (Whisper+T5) | 24.1 | Not recorded | 68fa7c0d1e209131cea8c22b | | SPECTRA | 21.96 | Not recorded | 68fa6fbc1e209131cea471d2 | | SPACE+WavLM+TripPy | 20.9 | Not recorded | 68fa6fbc1e209131cea471d2 | | Joint Speech and Text E2E DST | 20.7 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 19.8 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 18.9 | Not recorded | 692d2147d6b0dd334293190a | | Joint Speech and Text E2E DST | 18.7 | Not recorded | 692d2147d6b0dd334293190a | ### Full Spoken Context + Gemma2-9B-Instruct JGA: 45.52. Result ID: 699a89325698665ec41354e5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Setup":"potentially data-contaminated models"}. This record does not assert independent reproduction. ### Compressed Spoken Context + Gemma2-9B-Instruct JGA: 43.16. Result ID: 699a89325698665ec41354e4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Setup":"potentially data-contaminated models"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 43. Result ID: 6995de1a3badfaffe6e50c3e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"Gemma-3-12B-it","Training Text":"MW"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 42.6. Result ID: 6995de1a3badfaffe6e50c3d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"Gemma-3-12B-it","Training Text":"✗"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 42.2. Result ID: 6995de1a3badfaffe6e50c47. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"Gemma-3-12B-it","Training Text":"SW"}. This record does not assert independent reproduction. ### WavLM + conn. + Gemma2-9B-Instruct JGA: 42.17. Result ID: 699a89325698665ec41354e3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Setup":"potentially data-contaminated models"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 41. Result ID: 6995de1a3badfaffe6e50c3b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"Gemma-3-4B-it","Training Text":"✗"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 40.3. Result ID: 6995de1a3badfaffe6e50c3c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"Gemma-3-4B-it","Training Text":"MW"}. This record does not assert independent reproduction. ### Full Spoken Context JGA: 39.32. Result ID: 699a89325698665ec41354e2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Setup":"open data models"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 37.4. Result ID: 6995de1a3badfaffe6e50c45. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"Gemma-3-4B-it","Training Text":"SW"}. This record does not assert independent reproduction. ### Compressed Spoken Context JGA: 36.49. Result ID: 699a89325698665ec41354e1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Setup":"open data models"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 36.3. Result ID: 6995de1a3badfaffe6e50c3a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"Gemma-3-1B-it","Training Text":"MW"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 35.6. Result ID: 6995de1a3badfaffe6e50c39. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"Gemma-3-1B-it","Training Text":"✗"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 34.9. Result ID: 6995de1a3badfaffe6e50c41. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"OLMo-1B","Training Text":"MW"}. This record does not assert independent reproduction. ### WavLM + conn. + OLMo-1B JGA: 34.66. Result ID: 699a89325698665ec41354e0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Setup":"open data models"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 34.1. Result ID: 6995de1a3badfaffe6e50c40. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"OLMo-1B","Training Text":"✗"}. This record does not assert independent reproduction. ### OLMo-1B \[7\] JGA: 32.1. Result ID: 6995de1a3badfaffe6e50c3f. [Reporting paper]() Recorded evaluation configuration: {"Training Speech":"SW","LLM":"OLMo-1B","Training Text":"✗"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 30.9. Result ID: 6995de1a3badfaffe6e50c43. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"Gemma-3-1B-it","Training Text":"SW"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 29.3. Result ID: 6995de1a3badfaffe6e50c49. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"OLMo-1B","Training Text":"SW"}. This record does not assert independent reproduction. ### UBAR + GenWOZ JGA: 25.9. Result ID: 699a89325698665ec41354df. [Reporting paper]() Recorded evaluation configuration: {"Setup":"open data models"}. This record does not assert independent reproduction. ### SPACE+WavLMalign JGA: 25.65. Result ID: 699a89325698665ec41354dd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Setup":"open data models"}. This record does not assert independent reproduction. ### E2E (Whisper+T5) JGA: 24.1. Result ID: 699a89325698665ec41354de. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Setup":"open data models"}. This record does not assert independent reproduction. ### SPECTRA JGA: 21.96. Result ID: 6990e8fa3a6c3148490437aa. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"mode":"Fine-tuning"}. This record does not assert independent reproduction. ### SPACE+WavLM+TripPy JGA: 20.9. Result ID: 6990e8fa3a6c3148490437a9. [Reporting paper]() Recorded evaluation configuration: {"mode":"Fine-tuning"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 20.7. Result ID: 6995de1a3badfaffe6e50c46. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"Gemma-3-12B-it","Training Text":"✗"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 19.8. Result ID: 6995de1a3badfaffe6e50c44. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"Gemma-3-4B-it","Training Text":"✗"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 18.9. Result ID: 6995de1a3badfaffe6e50c42. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"Gemma-3-1B-it","Training Text":"✗"}. This record does not assert independent reproduction. ### Joint Speech and Text E2E DST JGA: 18.7. Result ID: 6995de1a3badfaffe6e50c48. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Speech":"MW","LLM":"OLMo-1B","Training Text":"✗"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.