Datasets:
Formats:
parquet
Languages:
English
Size:
10K - 100K
ArXiv:
Tags:
dialogue
hallucination
procedural-hallucination
task-oriented-dialogue
conversational-agents
llm-evaluation
License:
Card: train forks must not be used to train a method
Browse files
README.md
CHANGED
|
@@ -227,7 +227,7 @@ python scripts/download_data.py # -> datasets/dialfork/ (+ datasets/STAR/)
|
|
| 227 |
|
| 228 |
## Task 3 split
|
| 229 |
|
| 230 |
-
**Task 3** comes with an official `train`/`test` split, in the `task3_split` field. Task 3 is scored on `test`. Methods may learn only from the **conformant** dialogues of `train`, for example to learn the workflow: the paper's streaming detectors build their flow graph and n-gram model and calibrate their thresholds on them. The `train` forks and their labels must not be used to
|
| 231 |
|
| 232 |
| `task3_split` | source dialogues | dialogues |
|
| 233 |
|---|---:|---:|
|
|
|
|
| 227 |
|
| 228 |
## Task 3 split
|
| 229 |
|
| 230 |
+
**Task 3** comes with an official `train`/`test` split, in the `task3_split` field. Task 3 is scored on `test`. Methods may learn only from the **conformant** dialogues of `train`, for example to learn the workflow: the paper's streaming detectors build their flow graph and n-gram model and calibrate their thresholds on them. The `train` forks and their labels must not be used to train a method: they would let it exploit regularities of the construction (evenly spaced fork positions, a finite pool of injected behaviours) that do not exist in deployment, where failures are not known in advance. Supervised methods that use them should be reported separately, as they are not comparable. The split is over **source dialogues**: a conformant dialogue and its 16 forks are always on the same side, since a fork replays its source before the fork turn. Within each task, half of the source dialogues, chosen with a fixed seed, are in `test`:
|
| 231 |
|
| 232 |
| `task3_split` | source dialogues | dialogues |
|
| 233 |
|---|---:|---:|
|