Datasets:
Formats:
csv
Languages:
English
Size:
1K - 10K
Tags:
creative-evaluation
human-evaluation
preference-data
pairwise-comparison
evaluator-agreement
convergence-divergence
License:
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -168,7 +168,7 @@ Prompts were seeded from real creatives' work artifacts and lightly edited to st
|
|
| 168 |
|
| 169 |
## Limitations and scope
|
| 170 |
|
| 171 |
-
This is a focused benchmark, not a general capability leaderboard. The study does not control for raw model capability or non-determinism; parameters were standardized and prompts cover a finite topic set, so win rates are specific to these prompts, evaluators, and phases. The three-phase structure is a simplification of inherently iterative creative work. The evaluator pool is modest (31
|
| 172 |
|
| 173 |
## Working with Contra Labs
|
| 174 |
|
|
@@ -183,7 +183,7 @@ To commission an evaluation or dataset for your domain, reach out through [contr
|
|
| 183 |
|
| 184 |
## Provenance and consent
|
| 185 |
|
| 186 |
-
Evaluations were collected with consent from professional creatives
|
| 187 |
|
| 188 |
## Citation
|
| 189 |
|
|
|
|
| 168 |
|
| 169 |
## Limitations and scope
|
| 170 |
|
| 171 |
+
This is a focused benchmark, not a general capability leaderboard. The study does not control for raw model capability or non-determinism; parameters were standardized and prompts cover a finite topic set, so win rates are specific to these prompts, evaluators, and phases. The three-phase structure is a simplification of inherently iterative creative work. The evaluator pool is modest (31 designers, sourced from Contra's top-earning talent) and prompts were sampled once. Scalar ratings are populated only for the two code domains in this release (see note above). Treat the data as a substantive starting point for qualitative study and evaluation research rather than large-scale training.
|
| 172 |
|
| 173 |
## Working with Contra Labs
|
| 174 |
|
|
|
|
| 183 |
|
| 184 |
## Provenance and consent
|
| 185 |
|
| 186 |
+
Evaluations were collected with consent from professional creatives sourced from Contra's top-earning talent, selected by skillset and matched to the model category most relevant to their workflow. Evaluators are represented by anonymized numeric `user_id` values and a `user_core_skill` label only; no names or direct identifiers are included. Qualitative feedback was additionally stripped of personally identifiable information and model identities before processing. Model identities were blinded and output order randomized during evaluation to prevent brand bias.
|
| 187 |
|
| 188 |
## Citation
|
| 189 |
|