# Conversation Labeling on Cross-agent 80 labeled sessions (val)

Source: [Conversation Labeling on Cross-agent 80 labeled sessions (val)](<https://www.sota2.com/research/sota/conversation-labeling-on-cross-agent-80-labeled-sessions-val>)

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

Data updated: 2026-03-31T08:50:35.714Z

Evaluation of conversation labeling performance on the validation split of the Cross-agent 80 labeled sessions dataset, reporting metrics such as Overall Kappa, F1, and Kappa scores for Intent, Topic, and Follow-up actions.

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 | Conversation labeling |
| Dataset (including any stated split) | Cross-agent 80 labeled sessions (val) |
| Metric | Overall Kappa |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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

- [Overall Kappa](<https://www.sota2.com/research/sota/conversation-labeling-on-cross-agent-80-labeled-sessions-val.md?metric=Overall%20Kappa>): Higher is better; unit: SCALAR.

- [F1 Score](<https://www.sota2.com/research/sota/conversation-labeling-on-cross-agent-80-labeled-sessions-val.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

- [Intent Kappa](<https://www.sota2.com/research/sota/conversation-labeling-on-cross-agent-80-labeled-sessions-val.md?metric=Intent%20Kappa>): Higher is better; unit: SCALAR.

- [Topic Kappa](<https://www.sota2.com/research/sota/conversation-labeling-on-cross-agent-80-labeled-sessions-val.md?metric=Topic%20Kappa>): Higher is better; unit: SCALAR.

- [Follow-up Kappa](<https://www.sota2.com/research/sota/conversation-labeling-on-cross-agent-80-labeled-sessions-val.md?metric=Follow-up%20Kappa>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Overall Kappa | Result date | Report paper |
| --- | --- | --- | --- |
| Codex (OpenAI) | 0.93 | Not recorded | 69cb55020636e5bf4afc9ec7 |
| Gemini | 0.93 | Not recorded | 69cb55020636e5bf4afc9ec7 |
| Claude Code | 0.91 | Not recorded | 69cb55020636e5bf4afc9ec7 |
| Human-Human | 0.89 | Not recorded | 69cb55020636e5bf4afc9ec7 |

### Codex (OpenAI)

Overall Kappa: 0.93. Result ID: 69cb8ac747f34ce924b3f098.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55020636e5bf4afc9ec7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb55020636e5bf4afc9ec7.md>)

Recorded evaluation configuration: {"Best Ver.":"v7"}. This record does not assert independent reproduction.

### Gemini

Overall Kappa: 0.93. Result ID: 69cb8ac747f34ce924b3f099.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55020636e5bf4afc9ec7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb55020636e5bf4afc9ec7.md>)

Recorded evaluation configuration: {"Best Ver.":"v9"}. This record does not assert independent reproduction.

### Claude Code

Overall Kappa: 0.91. Result ID: 69cb8ac747f34ce924b3f09a.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55020636e5bf4afc9ec7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb55020636e5bf4afc9ec7.md>)

Recorded evaluation configuration: {"Best Ver.":"v5"}. This record does not assert independent reproduction.

### Human-Human

Overall Kappa: 0.89. Result ID: 69cb8ac747f34ce924b3f09b.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55020636e5bf4afc9ec7.md>)

Recorded evaluation configuration: {"Best Ver.":"–"}. This record does not assert independent reproduction.

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