Rename all to unbalanced, remove capped unbalanced config and review samples

Copy all/ to unbalanced/ (same LFS objects), remove the capped ten-class unbalanced config and the review samples, and update the card and Croissant metadata. The previous layout remains available at revision 364e3d31b5a96b0a524d0a5d18a46902f3f77020.

README.md

@@ -1,116 +1,6 @@
 ---
 license: cc-by-4.0
 dataset_info:
-- config_name: all
-  features:
-  - name: audio
-    dtype:
-      audio:
-        decode: false
-  - name: label
-    dtype:
-      class_label:
-        names:
-          '0': NSW_3
-          '1': NSW_2
-          '2': NSW_1
-          '3': SW_Dana
-          '4': SW_Luna
-          '5': SW_Nana
-          '6': SW_Neo
-          '7': SW_Nikita
-          '8': SW_Shy
-          '9': SW_Yosefa
-  - name: name
-    dtype: string
-  - name: onset
-    dtype: float32
-  - name: offset
-    dtype: float32
-  - name: duration
-    dtype: float32
-  - name: recording_duration
-    dtype: float32
-  - name: whistle_type
-    dtype: int64
-  - name: whistle_name
-    dtype: string
-  - name: f0_time
-    sequence: float32
-  - name: f0_hz
-    sequence: float32
-  - name: f0_conf
-    sequence: float32
-  - name: f0_ok
-    dtype: bool
-  - name: f0_bad_reason
-    dtype: string
-  - name: f0_spectrogram
-    dtype: image
-  - name: snr_db
-    dtype: float64
-  splits:
-  - name: train
-    num_examples: 5848
-  - name: validation
-    num_examples: 1253
-  - name: test
-    num_examples: 1253
-- config_name: all-review-sample
-  features:
-  - name: audio
-    dtype:
-      audio:
-        decode: false
-  - name: label
-    dtype:
-      class_label:
-        names:
-          '0': NSW_3
-          '1': NSW_2
-          '2': NSW_1
-          '3': SW_Dana
-          '4': SW_Luna
-          '5': SW_Nana
-          '6': SW_Neo
-          '7': SW_Nikita
-          '8': SW_Shy
-          '9': SW_Yosefa
-  - name: name
-    dtype: string
-  - name: onset
-    dtype: float32
-  - name: offset
-    dtype: float32
-  - name: duration
-    dtype: float32
-  - name: recording_duration
-    dtype: float32
-  - name: whistle_type
-    dtype: int64
-  - name: whistle_name
-    dtype: string
-  - name: f0_time
-    sequence: float32
-  - name: f0_hz
-    sequence: float32
-  - name: f0_conf
-    sequence: float32
-  - name: f0_ok
-    dtype: bool
-  - name: f0_bad_reason
-    dtype: string
-  - name: f0_spectrogram
-    dtype: image
-  - name: snr_db
-    dtype: float64
-  splits:
-  - name: train
-    num_examples: 336
-  - name: validation
-    num_examples: 72
-  - name: test
-    num_examples: 72
 - config_name: balanced
   features:
   - name: audio
@@ -160,55 +50,6 @@ dataset_info:
     num_examples: 450
   - name: test
     num_examples: 450
-- config_name: balanced-review-sample
-  features:
-  - name: audio
-    dtype:
-      audio:
-        decode: false
-  - name: label
-    dtype:
-      class_label:
-        names:
-          '0': NSW_1
-          '1': SW_Luna
-          '2': SW_Nana
-          '3': SW_Neo
-          '4': SW_Nikita
-          '5': SW_Yosefa
-  - name: name
-    dtype: string
-  - name: onset
-    dtype: float32
-  - name: offset
-    dtype: float32
-  - name: duration
-    dtype: float32
-  - name: recording_duration
-    dtype: float32
-  - name: whistle_type
-    dtype: int64
-  - name: whistle_name
-    dtype: string
-  - name: f0_time
-    sequence: float32
-  - name: f0_hz
-    sequence: float32
-  - name: f0_conf
-    sequence: float32
-  - name: f0_ok
-    dtype: bool
-  - name: f0_bad_reason
-    dtype: string
-  - name: f0_spectrogram
-    dtype: image
-  splits:
-  - name: train
-    num_examples: 336
-  - name: validation
-    num_examples: 72
-  - name: test
-    num_examples: 72
 - config_name: unbalanced
   features:
   - name: audio
@@ -255,66 +96,15 @@ dataset_info:
     dtype: string
   - name: f0_spectrogram
     dtype: image
+  - name: snr_db
+    dtype: float64
   splits:
   - name: train
-    num_examples: 2442
-  - name: validation
-    num_examples: 523
-  - name: test
-    num_examples: 523
-- config_name: unbalanced-review-sample
-  features:
-  - name: audio
-    dtype:
-      audio:
-        decode: false
-  - name: label
-    dtype:
-      class_label:
-        names:
-          '0': NSW_3
-          '1': NSW_2
-          '2': NSW_1
-          '3': SW_Dana
-          '4': SW_Luna
-          '5': SW_Nana
-          '6': SW_Neo
-          '7': SW_Nikita
-          '8': SW_Shy
-          '9': SW_Yosefa
-  - name: name
-    dtype: string
-  - name: onset
-    dtype: float32
-  - name: offset
-    dtype: float32
-  - name: duration
-    dtype: float32
-  - name: recording_duration
-    dtype: float32
-  - name: whistle_type
-    dtype: int64
-  - name: whistle_name
-    dtype: string
-  - name: f0_time
-    sequence: float32
-  - name: f0_hz
-    sequence: float32
-  - name: f0_conf
-    sequence: float32
-  - name: f0_ok
-    dtype: bool
-  - name: f0_bad_reason
-    dtype: string
-  - name: f0_spectrogram
-    dtype: image
-  splits:
-  - name: train
-    num_examples: 336
+    num_examples: 5848
   - name: validation
-    num_examples: 72
+    num_examples: 1253
   - name: test
-    num_examples: 72
+    num_examples: 1253
 task_categories:
 - audio-classification
 tags:
@@ -326,22 +116,6 @@ tags:
 - f0
 - spectrogram
 configs:
-- config_name: all
-  data_files:
-  - split: train
-    path: all/train-*
-  - split: validation
-    path: all/validation-*
-  - split: test
-    path: all/test-*
-- config_name: all-review-sample
-  data_files:
-  - split: train
-    path: all-review-sample/train-*
-  - split: validation
-    path: all-review-sample/validation-*
-  - split: test
-    path: all-review-sample/test-*
 - config_name: balanced
   data_files:
   - split: train
@@ -350,14 +124,6 @@ configs:
     path: balanced/validation-*
   - split: test
     path: balanced/test-*
-- config_name: balanced-review-sample
-  data_files:
-  - split: train
-    path: balanced-review-sample/train-*
-  - split: validation
-    path: balanced-review-sample/validation-*
-  - split: test
-    path: balanced-review-sample/test-*
 - config_name: unbalanced
   data_files:
   - split: train
@@ -366,14 +132,6 @@ configs:
     path: unbalanced/validation-*
   - split: test
     path: unbalanced/test-*
-- config_name: unbalanced-review-sample
-  data_files:
-  - split: train
-    path: unbalanced-review-sample/train-*
-  - split: validation
-    path: unbalanced-review-sample/validation-*
-  - split: test
-    path: unbalanced-review-sample/test-*
 ---
 
 # OpenWhistle Classification Finetuning Dataset
@@ -384,33 +142,36 @@ classification. It contains short whistle clips, whistle-level metadata,
 fundamental-frequency tracks, rendered F0 spectrograms, and integer class
 labels.
 
-The main reviewer-facing subset is the balanced `balanced` config. It contains
-six classes:
+The dataset has two configs:
 
-- `NSW_1` (`label=0`)
-- `SW_Luna` (`label=1`)
-- `SW_Nana` (`label=2`)
-- `SW_Neo` (`label=3`)
-- `SW_Nikita` (`label=4`)
-- `SW_Yosefa` (`label=5`)
+- `balanced`: the main six-class benchmark, with exactly 500 clips per class.
+- `unbalanced`: all 8,354 expert-annotated whistles across ten classes, with
+  the natural class imbalance.
 
-The full dataset is split by recording session, so no session appears in more
-than one of `train`, `validation`, or `test`. Smaller deterministic review
-configs are also provided so reviewers can inspect representative examples
-quickly without downloading the complete data first.
+Both configs are split by recording session, so no session appears in more
+than one of `train`, `validation`, or `test`.
 
 ## Dataset Contents
 
 - Hugging Face repo: `dolphinteam/OpenWhistle-Classification-Finetuning`
-- Main balanced config: `balanced`
-- Reviewer convenience config: `balanced-review-sample`
-- Public columns common to all configs: `audio`, `label`, `name`, `onset`, `offset`, `duration`,
-  `recording_duration`, `whistle_type`, `whistle_name`, `f0_time`, `f0_hz`,
-  `f0_conf`, `f0_ok`, `f0_bad_reason`, `f0_spectrogram`
-- The `all` and `all-review-sample` configs additionally include `snr_db`, the
-  estimated clip-level signal-to-noise ratio in dB.
+- Configs: `balanced` (main benchmark) and `unbalanced`
+- Audio: mono, 44.1 kHz
+- Public columns common to both configs: `audio`, `label`, `name`, `onset`,
+  `offset`, `duration`, `recording_duration`, `whistle_type`, `whistle_name`,
+  `f0_time`, `f0_hz`, `f0_conf`, `f0_ok`, `f0_bad_reason`, `f0_spectrogram`
+- The `unbalanced` config additionally includes `snr_db`, the estimated
+  clip-level signal-to-noise ratio in dB.
+
+## Balanced Config
 
-## Balanced Dataset Splits
+The `balanced` config contains six classes:
+
+- `NSW_1` (`label=0`)
+- `SW_Luna` (`label=1`)
+- `SW_Nana` (`label=2`)
+- `SW_Neo` (`label=3`)
+- `SW_Nikita` (`label=4`)
+- `SW_Yosefa` (`label=5`)
 
 | Split | Rows | NSW_1 | SW_Luna | SW_Nana | SW_Neo | SW_Nikita | SW_Yosefa | Sessions |
 | --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
@@ -422,93 +183,66 @@ quickly without downloading the complete data first.
 The split assignment was generated with seed `42` and exact class balancing.
 Session leakage checks found no overlap between any pair of splits.
 
-## Available Subsets
-
-The repository provides three full classification subsets and their smaller
-review counterparts:
-
-| Subset/config | Rows | Classes | Sessions | Purpose |
-| --- | ---: | ---: | ---: | --- |
-| `balanced` | 3,000 | 6 | 252 | Main balanced six-class finetuning dataset |
-| `unbalanced` | 3,488 | 10 | 258 | Ten-class finetuning dataset with capped rare classes |
-| `all` | 8,354 | 10 | 261 | Ten-class dataset preserving the full available class distribution |
-| `balanced-review-sample` | 480 | 6 | Same source split design | Small reviewer sample from `balanced` |
-| `unbalanced-review-sample` | 480 | 10 | Same source split design | Small reviewer sample from `unbalanced` |
-| `all-review-sample` | 480 | 10 | Same source split design | Small reviewer sample from `all` |
-
-The ten-class subsets use the following labels:
-
-- `NSW_3`
-- `NSW_2`
-- `NSW_1`
-- `SW_Dana`
-- `SW_Luna`
-- `SW_Nana`
-- `SW_Neo`
-- `SW_Nikita`
-- `SW_Shy`
-- `SW_Yosefa`
-
-The `balanced` subset keeps the six classes listed above and is the recommended
-starting point for reviewers and model finetuning. The `unbalanced` and `all`
-subsets expose the broader ten-class label space for additional analysis.
-
-## Review Samples
-
-Review samples are small deterministic subsets of the same public dataset. They
-were created only to make review and manual inspection easier. They are not a
-replacement for the full configs used for model development or reporting.
-
-### How The Review Samples Were Created
-
-All review samples were built after the session-disjoint train/validation/test
-splits were finalized. The review-sample scripts preserve the original split
-assignment: reviewer training examples come only from the original `train`
-split, reviewer validation examples only from `validation`, and reviewer test
-examples only from `test`.
-
-For `balanced-review-sample`, rows were sampled separately within each split and
-class. Each class group was shuffled deterministically with
-`numpy.default_rng(seed + split_index)` using seed `42`, then capped at 56 rows
-per class for `train` and 12 rows per class for both `validation` and `test`.
-This keeps the same 70/15/15 split ratio as the full `balanced` config while
-keeping every class equally represented.
-
-For `unbalanced-review-sample` and `all-review-sample`, the same deterministic
-shuffle was used, but the target rows were allocated proportionally to the
-source class distribution inside each split. This preserves the class imbalance
-of the larger source configs while keeping the review download small.
-
-### Review Sample Sizes
-
-| Config | Source config | Strategy | Train | Validation | Test | Total |
-| --- | --- | --- | ---: | ---: | ---: | ---: |
-| `balanced-review-sample` | `balanced` | Equal rows per class within each split | 336 | 72 | 72 | 480 |
-| `unbalanced-review-sample` | `unbalanced` | Proportional class distribution within each split | 336 | 72 | 72 | 480 |
-| `all-review-sample` | `all` | Proportional class distribution within each split | 336 | 72 | 72 | 480 |
-
-The reviewer-facing sample for the main balanced dataset is
-`balanced-review-sample`. The other review samples are included so each full
-subset has a matching small inspection subset.
+## Unbalanced Config
+
+The `unbalanced` config contains ten classes:
+
+- `NSW_3` (`label=0`)
+- `NSW_2` (`label=1`)
+- `NSW_1` (`label=2`)
+- `SW_Dana` (`label=3`)
+- `SW_Luna` (`label=4`)
+- `SW_Nana` (`label=5`)
+- `SW_Neo` (`label=6`)
+- `SW_Nikita` (`label=7`)
+- `SW_Shy` (`label=8`)
+- `SW_Yosefa` (`label=9`)
+
+Note that the integer labels differ from those of the `balanced` config.
+
+| Split | Rows | NSW_3 | NSW_2 | NSW_1 | SW_Dana | SW_Luna | SW_Nana | SW_Neo | SW_Nikita | SW_Shy | SW_Yosefa | Sessions |
+| --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
+| `train` | 5,848 | 31 | 19 | 460 | 235 | 2,054 | 365 | 1,567 | 622 | 57 | 438 | 145 |
+| `validation` | 1,253 | 7 | 4 | 99 | 50 | 440 | 78 | 336 | 133 | 12 | 94 | 54 |
+| `test` | 1,253 | 7 | 4 | 99 | 50 | 440 | 78 | 336 | 133 | 12 | 94 | 62 |
+| **Total** | **8,354** | **45** | **27** | **658** | **335** | **2,934** | **521** | **2,239** | **888** | **81** | **626** | **261** |
+
+No recording session appears in more than one split. The rare classes
+(`NSW_3`, `NSW_2`, `SW_Shy`) have few test examples, so per-class results for
+them should be interpreted with care.
 
 ## Loading The Data
 
 ```python
 from datasets import load_dataset
 
-full = load_dataset(
+balanced = load_dataset(
     "dolphinteam/OpenWhistle-Classification-Finetuning",
     "balanced",
 )
-review = load_dataset(
+unbalanced = load_dataset(
     "dolphinteam/OpenWhistle-Classification-Finetuning",
-    "balanced-review-sample",
+    "unbalanced",
 )
 ```
 
-Optional broader configs can be loaded by passing `"unbalanced"` or `"all"` as
-the second `load_dataset` argument. Their corresponding review configs are
-`"unbalanced-review-sample"` and `"all-review-sample"`.
+## Earlier Revisions
+
+In earlier revisions of this repository, the current `unbalanced` config was
+named `all`; its data files are unchanged, so results reported for `all`
+correspond to `unbalanced`. Those revisions also contained a smaller ten-class
+config with capped class sizes, then named `unbalanced` (3,488 clips), and
+`*-review-sample` configs for inspection. They remain available by revision:
+
+```python
+from datasets import load_dataset
+
+previous = load_dataset(
+    "dolphinteam/OpenWhistle-Classification-Finetuning",
+    "all",
+    revision="364e3d31b5a96b0a524d0a5d18a46902f3f77020",
+)
+```
 
 ## License
 

balanced-review-sample/train-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:de3cc9645b4e40f6665392a767026d33598ea7382245b208dbb4931fd991ec10
-size 136612956

balanced-review-sample/validation-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:70f81bc564efdc33fc4849ea7fa1de75da5f44837206671546a09fd6acfae1cc
-size 28952403

croissant_rai_OpenWhistle-Classification-Finetuning.json

@@ -57,9 +57,9 @@
     "dolphinteam/OpenWhistle-Classification-Finetuning",
     "OpenWhistle-1.0-Classification-Finetuning"
   ],
-  "description": "OpenWhistle 1.0 Classification Finetuning is a public dolphin whistle-type classification dataset. It contains short whistle clips, expert-refined whistle labels, whistle-level timing metadata, fundamental-frequency tracks, rendered F0 spectrograms, and train/validation/test splits. The main reviewer-facing balanced configuration contains 3,000 examples across six classes with exact class balancing and session-disjoint splits; broader all and unbalanced configurations expose ten whistle categories and the naturally imbalanced class distribution.",
+  "description": "OpenWhistle 1.0 Classification Finetuning is a public dolphin whistle-type classification dataset. It contains short mono whistle clips sampled at 44.1 kHz, expert-refined whistle labels, whistle-level timing metadata, fundamental-frequency tracks, rendered F0 spectrograms, and train/validation/test splits. The main balanced configuration contains 3,000 examples across six classes with exact class balancing and session-disjoint splits; the unbalanced configuration contains all 8,354 annotated whistles across ten classes with the naturally imbalanced class distribution.",
   "url": "https://huggingface.co/datasets/dolphinteam/OpenWhistle-Classification-Finetuning",
-  "citeAs": "Faadil Mustun, Chiara Semenzin, Roberto Dessi, Pablo Robin Guerrero, Pierre Orhan, Alexis Emanuelli, Emanuele Rossi, Yair Lakretz, Gonzalo de Polavieja, and German Sumbre. OpenWhistle: A Large-Scale Longitudinal Dataset and Benchmark of Bottlenose Dolphin Vocalizations. NeurIPS 2026 Datasets & Evaluations Track (Spotlight), arXiv:2609.34839. https://arxiv.org/abs/2609.34839",
+  "citeAs": "@misc{mustun2026openwhistle,\n  title         = {OpenWhistle: A Large-Scale Longitudinal Dataset and Benchmark of Bottlenose Dolphin Vocalizations},\n  author        = {Mustun, Faadil and Semenzin, Chiara and Dessi, Roberto and Robin Guerrero, Pablo and Orhan, Pierre and Emanuelli, Alexis and Rossi, Emanuele and Lakretz, Yair and de Polavieja, Gonzalo and Sumbre, German},\n  year          = {2026},\n  eprint        = {2609.34839},\n  archivePrefix = {arXiv},\n  primaryClass  = {cs.CL},\n  url           = {https://arxiv.org/abs/2609.34839}\n}",
   "creator": [
     {
       "@type": "Person",
@@ -108,7 +108,7 @@
     "url": "https://huggingface.co/dolphinteam"
   },
   "datePublished": "2026-04-20",
-  "dateModified": "2026-09-29",
+  "dateModified": "2026-11-06",
   "license": "https://creativecommons.org/licenses/by/4.0/",
   "version": "1.0.0",
   "identifier": "https://huggingface.co/datasets/dolphinteam/OpenWhistle-Classification-Finetuning",
@@ -148,1212 +148,102 @@
     },
     {
       "@type": "cr:FileSet",
-      "@id": "parquet-files-for-config-all",
-      "containedIn": {
-        "@id": "repo"
-      },
-      "encodingFormat": "application/x-parquet",
-      "includes": "all/*/*.parquet"
-    },
-    {
-      "@type": "cr:FileSet",
-      "@id": "parquet-files-for-config-all-review-sample",
-      "containedIn": {
-        "@id": "repo"
-      },
-      "encodingFormat": "application/x-parquet",
-      "includes": "all-review-sample/*/*.parquet"
-    },
-    {
-      "@type": "cr:FileSet",
-      "@id": "parquet-files-for-config-balanced",
-      "containedIn": {
-        "@id": "repo"
-      },
-      "encodingFormat": "application/x-parquet",
-      "includes": "balanced/*/*.parquet"
-    },
-    {
-      "@type": "cr:FileSet",
-      "@id": "parquet-files-for-config-balanced-review-sample",
-      "containedIn": {
-        "@id": "repo"
-      },
-      "encodingFormat": "application/x-parquet",
-      "includes": "balanced-review-sample/*/*.parquet"
-    },
-    {
-      "@type": "cr:FileSet",
-      "@id": "parquet-files-for-config-unbalanced",
-      "containedIn": {
-        "@id": "repo"
-      },
-      "encodingFormat": "application/x-parquet",
-      "includes": "unbalanced/*/*.parquet"
-    },
-    {
-      "@type": "cr:FileSet",
-      "@id": "parquet-files-for-config-unbalanced-review-sample",
-      "containedIn": {
-        "@id": "repo"
-      },
-      "encodingFormat": "application/x-parquet",
-      "includes": "unbalanced-review-sample/*/*.parquet"
-    }
-  ],
-  "recordSet": [
-    {
-      "@type": "cr:RecordSet",
-      "@id": "all_splits",
-      "name": "all_splits",
-      "description": "Splits for the all config.",
-      "dataType": "cr:Split",
-      "key": {
-        "@id": "all_splits/split_name"
-      },
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "all_splits/split_name",
-          "dataType": "sc:Text"
-        }
-      ],
-      "data": [
-        {
-          "all_splits/split_name": "train"
-        },
-        {
-          "all_splits/split_name": "validation"
-        },
-        {
-          "all_splits/split_name": "test"
-        }
-      ]
-    },
-    {
-      "@type": "cr:RecordSet",
-      "@id": "all",
-      "name": "all",
-      "description": "Full classification configuration exposing the broader ten-class whistle label space with train, validation, and test splits.",
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "all/split",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "fileProperty": "fullpath"
-            },
-            "transform": {
-              "regex": "all/(?:partial-)?(train|validation|test)/.+parquet$"
-            }
-          },
-          "references": {
-            "field": {
-              "@id": "all_splits/split_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/audio",
-          "dataType": "sc:AudioObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "audio"
-            }
-          },
-          "samplingRate": 96000
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/label",
-          "dataType": "sc:Integer",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "label"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/onset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "onset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/offset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "offset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/recording_duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "recording_duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/whistle_type",
-          "dataType": "cr:Int64",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "whistle_type"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/whistle_name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "whistle_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/f0_time",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "f0_time"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/f0_hz",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "f0_hz"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/f0_conf",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "f0_conf"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/f0_ok",
-          "dataType": "sc:Boolean",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "f0_ok"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/f0_bad_reason",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "f0_bad_reason"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/f0_spectrogram",
-          "dataType": "sc:ImageObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "f0_spectrogram"
-            },
-            "transform": {
-              "jsonPath": "bytes"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all/snr_db",
-          "dataType": "cr:Float64",
-          "description": "Estimated clip-level signal-to-noise ratio in dB.",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all"
-            },
-            "extract": {
-              "column": "snr_db"
-            }
-          }
-        }
-      ]
-    },
-    {
-      "@type": "cr:RecordSet",
-      "@id": "all-review-sample_splits",
-      "name": "all-review-sample_splits",
-      "description": "Splits for the all-review-sample config.",
-      "dataType": "cr:Split",
-      "key": {
-        "@id": "all-review-sample_splits/split_name"
-      },
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample_splits/split_name",
-          "dataType": "sc:Text"
-        }
-      ],
-      "data": [
-        {
-          "all-review-sample_splits/split_name": "train"
-        },
-        {
-          "all-review-sample_splits/split_name": "validation"
-        },
-        {
-          "all-review-sample_splits/split_name": "test"
-        }
-      ]
-    },
-    {
-      "@type": "cr:RecordSet",
-      "@id": "all-review-sample",
-      "name": "all-review-sample",
-      "description": "Deterministic all-configuration review sample for reviewer inspection.",
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/split",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "fileProperty": "fullpath"
-            },
-            "transform": {
-              "regex": "all\\-review\\-sample/(?:partial-)?(train|validation|test)/.+parquet$"
-            }
-          },
-          "references": {
-            "field": {
-              "@id": "all-review-sample_splits/split_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/audio",
-          "dataType": "sc:AudioObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "audio"
-            }
-          },
-          "samplingRate": 96000
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/label",
-          "dataType": "sc:Integer",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "label"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/onset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "onset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/offset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "offset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/recording_duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "recording_duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/whistle_type",
-          "dataType": "cr:Int64",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "whistle_type"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/whistle_name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "whistle_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/f0_time",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "f0_time"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/f0_hz",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "f0_hz"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/f0_conf",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "f0_conf"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/f0_ok",
-          "dataType": "sc:Boolean",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "f0_ok"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/f0_bad_reason",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "f0_bad_reason"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/f0_spectrogram",
-          "dataType": "sc:ImageObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "f0_spectrogram"
-            },
-            "transform": {
-              "jsonPath": "bytes"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "all-review-sample/snr_db",
-          "dataType": "cr:Float64",
-          "description": "Estimated clip-level signal-to-noise ratio in dB.",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-all-review-sample"
-            },
-            "extract": {
-              "column": "snr_db"
-            }
-          }
-        }
-      ]
-    },
-    {
-      "@type": "cr:RecordSet",
-      "@id": "balanced_splits",
-      "name": "balanced_splits",
-      "description": "Splits for the balanced config.",
-      "dataType": "cr:Split",
-      "key": {
-        "@id": "balanced_splits/split_name"
-      },
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "balanced_splits/split_name",
-          "dataType": "sc:Text"
-        }
-      ],
-      "data": [
-        {
-          "balanced_splits/split_name": "train"
-        },
-        {
-          "balanced_splits/split_name": "validation"
-        },
-        {
-          "balanced_splits/split_name": "test"
-        }
-      ]
-    },
-    {
-      "@type": "cr:RecordSet",
-      "@id": "balanced",
-      "name": "balanced",
-      "description": "Balanced reviewer-facing six-class classification benchmark with session-disjoint train, validation, and test splits.",
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/split",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "fileProperty": "fullpath"
-            },
-            "transform": {
-              "regex": "balanced/(?:partial-)?(train|validation|test)/.+parquet$"
-            }
-          },
-          "references": {
-            "field": {
-              "@id": "balanced_splits/split_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/audio",
-          "dataType": "sc:AudioObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "audio"
-            }
-          },
-          "samplingRate": 96000
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/label",
-          "dataType": "sc:Integer",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "label"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/onset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "onset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/offset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "offset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/recording_duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "recording_duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/whistle_type",
-          "dataType": "cr:Int64",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "whistle_type"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/whistle_name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "whistle_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/f0_time",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "f0_time"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/f0_hz",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "f0_hz"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/f0_conf",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "f0_conf"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/f0_ok",
-          "dataType": "sc:Boolean",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "f0_ok"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/f0_bad_reason",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "f0_bad_reason"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced/f0_spectrogram",
-          "dataType": "sc:ImageObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced"
-            },
-            "extract": {
-              "column": "f0_spectrogram"
-            },
-            "transform": {
-              "jsonPath": "bytes"
-            }
-          }
-        }
-      ]
-    },
-    {
-      "@type": "cr:RecordSet",
-      "@id": "balanced-review-sample_splits",
-      "name": "balanced-review-sample_splits",
-      "description": "Splits for the balanced-review-sample config.",
-      "dataType": "cr:Split",
-      "key": {
-        "@id": "balanced-review-sample_splits/split_name"
-      },
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample_splits/split_name",
-          "dataType": "sc:Text"
-        }
-      ],
-      "data": [
-        {
-          "balanced-review-sample_splits/split_name": "train"
-        },
-        {
-          "balanced-review-sample_splits/split_name": "validation"
-        },
-        {
-          "balanced-review-sample_splits/split_name": "test"
-        }
-      ]
-    },
-    {
-      "@type": "cr:RecordSet",
-      "@id": "balanced-review-sample",
-      "name": "balanced-review-sample",
-      "description": "Deterministic balanced review sample for reviewer inspection of the six-class benchmark.",
-      "field": [
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/split",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "fileProperty": "fullpath"
-            },
-            "transform": {
-              "regex": "balanced\\-review\\-sample/(?:partial-)?(train|validation|test)/.+parquet$"
-            }
-          },
-          "references": {
-            "field": {
-              "@id": "balanced-review-sample_splits/split_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/audio",
-          "dataType": "sc:AudioObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "audio"
-            }
-          },
-          "samplingRate": 96000
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/label",
-          "dataType": "sc:Integer",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "label"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/onset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "onset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/offset",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "offset"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/recording_duration",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "recording_duration"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/whistle_type",
-          "dataType": "cr:Int64",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "whistle_type"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/whistle_name",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "whistle_name"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/f0_time",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "f0_time"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/f0_hz",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "f0_hz"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/f0_conf",
-          "dataType": "cr:Float32",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "f0_conf"
-            }
-          },
-          "isArray": true,
-          "arrayShape": "-1"
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/f0_ok",
-          "dataType": "sc:Boolean",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "f0_ok"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/f0_bad_reason",
-          "dataType": "sc:Text",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "f0_bad_reason"
-            }
-          }
-        },
-        {
-          "@type": "cr:Field",
-          "@id": "balanced-review-sample/f0_spectrogram",
-          "dataType": "sc:ImageObject",
-          "source": {
-            "fileSet": {
-              "@id": "parquet-files-for-config-balanced-review-sample"
-            },
-            "extract": {
-              "column": "f0_spectrogram"
-            },
-            "transform": {
-              "jsonPath": "bytes"
-            }
-          }
-        }
-      ]
+      "@id": "parquet-files-for-config-balanced",
+      "containedIn": {
+        "@id": "repo"
+      },
+      "encodingFormat": "application/x-parquet",
+      "includes": "balanced/*/*.parquet"
     },
+    {
+      "@type": "cr:FileSet",
+      "@id": "parquet-files-for-config-unbalanced",
+      "containedIn": {
+        "@id": "repo"
+      },
+      "encodingFormat": "application/x-parquet",
+      "includes": "unbalanced/*/*.parquet"
+    }
+  ],
+  "recordSet": [
     {
       "@type": "cr:RecordSet",
-      "@id": "unbalanced_splits",
-      "name": "unbalanced_splits",
-      "description": "Splits for the unbalanced config.",
+      "@id": "balanced_splits",
+      "name": "balanced_splits",
+      "description": "Splits for the balanced config.",
       "dataType": "cr:Split",
       "key": {
-        "@id": "unbalanced_splits/split_name"
+        "@id": "balanced_splits/split_name"
       },
       "field": [
         {
           "@type": "cr:Field",
-          "@id": "unbalanced_splits/split_name",
+          "@id": "balanced_splits/split_name",
           "dataType": "sc:Text"
         }
       ],
       "data": [
         {
-          "unbalanced_splits/split_name": "train"
+          "balanced_splits/split_name": "train"
         },
         {
-          "unbalanced_splits/split_name": "validation"
+          "balanced_splits/split_name": "validation"
         },
         {
-          "unbalanced_splits/split_name": "test"
+          "balanced_splits/split_name": "test"
         }
       ]
     },
     {
       "@type": "cr:RecordSet",
-      "@id": "unbalanced",
-      "name": "unbalanced",
-      "description": "Naturally imbalanced classification configuration exposing ten whistle classes with train, validation, and test splits.",
+      "@id": "balanced",
+      "name": "balanced",
+      "description": "Main six-class balanced classification benchmark: 3,000 clips (2,100 train, 450 validation, 450 test; 350, 75, and 75 per class) from 252 recording sessions, with session-disjoint splits.",
       "field": [
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/split",
+          "@id": "balanced/split",
           "dataType": "sc:Text",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "fileProperty": "fullpath"
             },
             "transform": {
-              "regex": "unbalanced/(?:partial-)?(train|validation|test)/.+parquet$"
+              "regex": "balanced/(?:partial-)?(train|validation|test)/.+parquet$"
             }
           },
           "references": {
             "field": {
-              "@id": "unbalanced_splits/split_name"
+              "@id": "balanced_splits/split_name"
             }
           }
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/audio",
+          "@id": "balanced/audio",
           "dataType": "sc:AudioObject",
+          "description": "Mono whistle clip sampled at 44.1 kHz.",
+          "samplingRate": 44100,
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "audio"
             }
-          },
-          "samplingRate": 96000
+          }
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/label",
+          "@id": "balanced/label",
           "dataType": "sc:Integer",
+          "description": "Integer class label: 0=NSW_1, 1=SW_Luna, 2=SW_Nana, 3=SW_Neo, 4=SW_Nikita, 5=SW_Yosefa.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "label"
@@ -1362,11 +252,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/name",
+          "@id": "balanced/name",
           "dataType": "sc:Text",
+          "description": "Recording session identifier.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "name"
@@ -1375,11 +266,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/onset",
+          "@id": "balanced/onset",
           "dataType": "cr:Float32",
+          "description": "Whistle onset in seconds within the source recording.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "onset"
@@ -1388,11 +280,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/offset",
+          "@id": "balanced/offset",
           "dataType": "cr:Float32",
+          "description": "Whistle offset in seconds within the source recording.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "offset"
@@ -1401,11 +294,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/duration",
+          "@id": "balanced/duration",
           "dataType": "cr:Float32",
+          "description": "Whistle clip duration in seconds.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "duration"
@@ -1414,11 +308,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/recording_duration",
+          "@id": "balanced/recording_duration",
           "dataType": "cr:Float32",
+          "description": "Duration of the source recording in seconds.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "recording_duration"
@@ -1427,11 +322,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/whistle_type",
+          "@id": "balanced/whistle_type",
           "dataType": "cr:Int64",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "whistle_type"
@@ -1440,11 +335,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/whistle_name",
+          "@id": "balanced/whistle_name",
           "dataType": "sc:Text",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "whistle_name"
@@ -1453,11 +348,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/f0_time",
+          "@id": "balanced/f0_time",
           "dataType": "cr:Float32",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "f0_time"
@@ -1468,11 +363,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/f0_hz",
+          "@id": "balanced/f0_hz",
           "dataType": "cr:Float32",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "f0_hz"
@@ -1483,11 +378,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/f0_conf",
+          "@id": "balanced/f0_conf",
           "dataType": "cr:Float32",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "f0_conf"
@@ -1498,11 +393,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/f0_ok",
+          "@id": "balanced/f0_ok",
           "dataType": "sc:Boolean",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "f0_ok"
@@ -1511,11 +406,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/f0_bad_reason",
+          "@id": "balanced/f0_bad_reason",
           "dataType": "sc:Text",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "f0_bad_reason"
@@ -1524,11 +419,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced/f0_spectrogram",
+          "@id": "balanced/f0_spectrogram",
           "dataType": "sc:ImageObject",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced"
+              "@id": "parquet-files-for-config-balanced"
             },
             "extract": {
               "column": "f0_spectrogram"
@@ -1542,80 +437,82 @@
     },
     {
       "@type": "cr:RecordSet",
-      "@id": "unbalanced-review-sample_splits",
-      "name": "unbalanced-review-sample_splits",
-      "description": "Splits for the unbalanced-review-sample config.",
+      "@id": "unbalanced_splits",
+      "name": "unbalanced_splits",
+      "description": "Splits for the unbalanced config.",
       "dataType": "cr:Split",
       "key": {
-        "@id": "unbalanced-review-sample_splits/split_name"
+        "@id": "unbalanced_splits/split_name"
       },
       "field": [
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample_splits/split_name",
+          "@id": "unbalanced_splits/split_name",
           "dataType": "sc:Text"
         }
       ],
       "data": [
         {
-          "unbalanced-review-sample_splits/split_name": "train"
+          "unbalanced_splits/split_name": "train"
         },
         {
-          "unbalanced-review-sample_splits/split_name": "validation"
+          "unbalanced_splits/split_name": "validation"
         },
         {
-          "unbalanced-review-sample_splits/split_name": "test"
+          "unbalanced_splits/split_name": "test"
         }
       ]
     },
     {
       "@type": "cr:RecordSet",
-      "@id": "unbalanced-review-sample",
-      "name": "unbalanced-review-sample",
-      "description": "Deterministic unbalanced-configuration review sample for reviewer inspection.",
+      "@id": "unbalanced",
+      "name": "unbalanced",
+      "description": "Ten-class classification configuration with the full, naturally imbalanced class distribution: 8,354 clips (5,848 train, 1,253 validation, 1,253 test) from 261 recording sessions, with session-disjoint splits. Includes the clip-level snr_db column.",
       "field": [
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/split",
+          "@id": "unbalanced/split",
           "dataType": "sc:Text",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "fileProperty": "fullpath"
             },
             "transform": {
-              "regex": "unbalanced\\-review\\-sample/(?:partial-)?(train|validation|test)/.+parquet$"
+              "regex": "unbalanced/(?:partial-)?(train|validation|test)/.+parquet$"
             }
           },
           "references": {
             "field": {
-              "@id": "unbalanced-review-sample_splits/split_name"
+              "@id": "unbalanced_splits/split_name"
             }
           }
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/audio",
+          "@id": "unbalanced/audio",
           "dataType": "sc:AudioObject",
+          "description": "Mono whistle clip sampled at 44.1 kHz.",
+          "samplingRate": 44100,
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "audio"
             }
-          },
-          "samplingRate": 96000
+          }
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/label",
+          "@id": "unbalanced/label",
           "dataType": "sc:Integer",
+          "description": "Integer class label: 0=NSW_3, 1=NSW_2, 2=NSW_1, 3=SW_Dana, 4=SW_Luna, 5=SW_Nana, 6=SW_Neo, 7=SW_Nikita, 8=SW_Shy, 9=SW_Yosefa.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "label"
@@ -1624,11 +521,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/name",
+          "@id": "unbalanced/name",
           "dataType": "sc:Text",
+          "description": "Recording session identifier.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "name"
@@ -1637,11 +535,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/onset",
+          "@id": "unbalanced/onset",
           "dataType": "cr:Float32",
+          "description": "Whistle onset in seconds within the source recording.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "onset"
@@ -1650,11 +549,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/offset",
+          "@id": "unbalanced/offset",
           "dataType": "cr:Float32",
+          "description": "Whistle offset in seconds within the source recording.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "offset"
@@ -1663,11 +563,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/duration",
+          "@id": "unbalanced/duration",
           "dataType": "cr:Float32",
+          "description": "Whistle clip duration in seconds.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "duration"
@@ -1676,11 +577,12 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/recording_duration",
+          "@id": "unbalanced/recording_duration",
           "dataType": "cr:Float32",
+          "description": "Duration of the source recording in seconds.",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "recording_duration"
@@ -1689,11 +591,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/whistle_type",
+          "@id": "unbalanced/whistle_type",
           "dataType": "cr:Int64",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "whistle_type"
@@ -1702,11 +604,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/whistle_name",
+          "@id": "unbalanced/whistle_name",
           "dataType": "sc:Text",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "whistle_name"
@@ -1715,11 +617,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/f0_time",
+          "@id": "unbalanced/f0_time",
           "dataType": "cr:Float32",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "f0_time"
@@ -1730,11 +632,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/f0_hz",
+          "@id": "unbalanced/f0_hz",
           "dataType": "cr:Float32",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "f0_hz"
@@ -1745,11 +647,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/f0_conf",
+          "@id": "unbalanced/f0_conf",
           "dataType": "cr:Float32",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "f0_conf"
@@ -1760,11 +662,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/f0_ok",
+          "@id": "unbalanced/f0_ok",
           "dataType": "sc:Boolean",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "f0_ok"
@@ -1773,11 +675,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/f0_bad_reason",
+          "@id": "unbalanced/f0_bad_reason",
           "dataType": "sc:Text",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "f0_bad_reason"
@@ -1786,11 +688,11 @@
         },
         {
           "@type": "cr:Field",
-          "@id": "unbalanced-review-sample/f0_spectrogram",
+          "@id": "unbalanced/f0_spectrogram",
           "dataType": "sc:ImageObject",
           "source": {
             "fileSet": {
-              "@id": "parquet-files-for-config-unbalanced-review-sample"
+              "@id": "parquet-files-for-config-unbalanced"
             },
             "extract": {
               "column": "f0_spectrogram"
@@ -1799,6 +701,20 @@
               "jsonPath": "bytes"
             }
           }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "unbalanced/snr_db",
+          "dataType": "cr:Float64",
+          "description": "Estimated clip-level signal-to-noise ratio in dB.",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-unbalanced"
+            },
+            "extract": {
+              "column": "snr_db"
+            }
+          }
         }
       ]
     }
@@ -1810,27 +726,49 @@
     "Manual Human Curator",
     "Software Collection"
   ],
-  "rai:dataCollectionRawData": "Raw data consisted of underwater acoustic recordings captured by fixed and hidden hydrophones. The released classification dataset contains short whistle clips and metadata fields: audio, label, name, onset, offset, duration, recording_duration, whistle_type, whistle_name, f0_time, f0_hz, f0_conf, f0_ok, f0_bad_reason, and f0_spectrogram. The all and all-review-sample configurations additionally include snr_db, an estimated clip-level signal-to-noise ratio in dB.",
-  "rai:dataCollectionMissingData": "The public dataset does not include per-example video context, exact hydrophone coordinates, environmental conditions, full behavioral context, or human-identifying information. The balanced configuration intentionally excludes rare whistle categories to support reliable six-class evaluation, while all and unbalanced configurations retain a broader ten-class label space with imbalanced classes.",
+  "rai:dataCollectionRawData": "Raw data consisted of underwater acoustic recordings captured by fixed and hidden hydrophones. The released classification dataset contains short mono whistle clips sampled at 44.1 kHz and metadata fields: audio, label, name, onset, offset, duration, recording_duration, whistle_type, whistle_name, f0_time, f0_hz, f0_conf, f0_ok, f0_bad_reason, and f0_spectrogram. The unbalanced configuration additionally includes snr_db, an estimated clip-level signal-to-noise ratio in dB.",
+  "rai:dataCollectionMissingData": "The public dataset does not include per-example video context, exact hydrophone coordinates, environmental conditions, full behavioral context, or human-identifying information. The balanced configuration intentionally excludes rare whistle categories to support reliable six-class evaluation, while the unbalanced configuration retains the broader ten-class label space with imbalanced classes.",
   "rai:dataPreprocessingProtocol": [
-    "Whistle clips were extracted from OpenWhistle recordings and packaged as Hugging Face parquet datasets with audio, timing metadata, F0 contours, F0 confidence values, rendered F0 spectrogram images, and, for the all and all-review-sample configurations, clip-level snr_db estimates.",
-    "The main balanced configuration uses six classes and exact class balancing with 2,100 train, 450 validation, and 450 test examples. Splits are session-disjoint, so no recording session appears in more than one split.",
-    "The unbalanced and all configurations expose ten whistle classes for additional analysis. Review-sample configurations were generated deterministically after split creation and are intended for manual inspection, not reporting."
+    "Whistle clips were extracted from OpenWhistle recordings and packaged as Hugging Face parquet datasets with mono 44.1 kHz audio, timing metadata, F0 contours, F0 confidence values, rendered F0 spectrogram images, and, for the unbalanced configuration, clip-level snr_db estimates.",
+    "The main balanced configuration uses six classes and exact class balancing with 2,100 train, 450 validation, and 450 test examples (350, 75, and 75 per class) from 161, 54, and 37 recording sessions. Splits are session-disjoint, so no recording session appears in more than one split.",
+    "The unbalanced configuration exposes ten whistle classes for additional analysis: 8,354 clips (5,848 train, 1,253 validation, 1,253 test) from 261 recording sessions, with session-disjoint splits."
   ],
   "rai:dataAnnotationProtocol": "Whistle categories were obtained through an annotation pipeline combining automated contour-based grouping with expert refinement. Detected whistle segments were categorized into known signature-whistle and non-signature-whistle types, then manually refined by expert annotators through visual inspection of spectrograms to correct misclassifications and resolve ambiguous cases.",
   "rai:annotationsPerItem": "Each released example has one whistle-type class label. The label corresponds to either a signature whistle class or a non-signature whistle class, depending on the configuration.",
   "rai:machineAnnotationTools": [
     "ARTwarp and dynamic time warping were used in the annotation pipeline to cluster whistle contours before expert refinement.",
-    "F0 extraction and spectrogram rendering software were used to create f0_time, f0_hz, f0_conf, f0_ok, f0_bad_reason, and f0_spectrogram fields. Signal-to-noise estimation software was used to create snr_db for the all and all-review-sample configurations.",
+    "F0 extraction and spectrogram rendering software were used to create f0_time, f0_hz, f0_conf, f0_ok, f0_bad_reason, and f0_spectrogram fields. Signal-to-noise estimation software was used to create snr_db for the unbalanced configuration.",
     "Hugging Face Datasets tooling was used to package and publish the dataset as parquet configurations."
   ],
-  "rai:dataUseCases": "[\"Construct represented: short dolphin whistle clips with expert-refined whistle-type labels, timing metadata, F0 tracks, and F0 spectrograms.\", \"Validated use case: training, validation, and testing of dolphin whistle-type classifiers on the balanced six-class OpenWhistle benchmark with session-disjoint train, validation, and test splits.\", \"Validated use case: linear probing or finetuning of audio representation models for fine-grained intra-species dolphin whistle discrimination.\", \"Supported exploratory use case: analysis of ten-class whistle categories using the all and unbalanced configurations, with explicit attention to class imbalance.\", \"Supported reviewer use case: manual inspection through deterministic review-sample configurations.\", \"Not validated use case: inferring behavioral intent, communicative meaning, animal welfare state, individual identity beyond the released label definitions, or deployment performance in other sites without additional biological validation.\"]",
-  "rai:dataLimitations": "[\"The balanced configuration is designed for six-class whistle-type classification and should not be treated as a complete representation of all dolphin whistle categories or all individuals at the site.\", \"The all and unbalanced configurations expose ten classes but contain strong class imbalance, including rare whistle types that may not support reliable standalone evaluation.\", \"The dataset comes from a semi-natural site with a stable pod of known dolphins, so results may not generalize to fully wild populations, other species, other recording devices, or different acoustic environments without validation.\", \"Labels describe whistle categories, not full behavioral intent, communicative meaning, animal state, or ecological outcomes.\", \"Review-sample configurations are for inspection only and should not replace the full configurations for model development or reporting.\", \"Session-disjoint splits reduce session leakage, but downstream users should still avoid mixing sessions or configurations in ways that invalidate the benchmark protocol.\"]",
-  "rai:dataBiases": "[\"The dataset reflects recordings from Dolphin Reef, Eilat and may overrepresent acoustic, social, and environmental conditions at that semi-natural site.\", \"The naturally available whistle repertoire is imbalanced: some signature whistle and non-signature whistle categories are much more frequent than others.\", \"The main balanced configuration intentionally subsamples classes to equalize six labels, which improves benchmark comparability but no longer reflects natural production frequencies.\", \"Expert-refined labels depend on the available template contours, annotation pipeline, and annotator judgments, so ambiguous or low-quality whistles may be underrepresented.\", \"Examples with usable F0 tracks or clearer spectrograms may be easier to analyze and may not fully represent all noisy, overlapping, or low-SNR acoustic conditions.\"]",
-  "rai:personalSensitiveInformation": "[\"No human subjects are involved in the dataset. Public columns contain animal audio clips, timing fields, whistle labels, F0 tracks, F0 spectrograms, and recording/session metadata.\", \"The data consists of underwater animal acoustic recordings. Users should nevertheless treat incidental non-target sounds as possible and avoid attempts to infer people, vessels, precise locations, or sensitive field-site details from the audio.\", \"Whistle labels may be associated with known dolphin signature-whistle categories, but the public dataset does not include sensitive human information.\"]",
+  "rai:dataUseCases": [
+    "Construct represented: short dolphin whistle clips with expert-refined whistle-type labels, timing metadata, F0 tracks, and F0 spectrograms.",
+    "Validated use case: training, validation, and testing of dolphin whistle-type classifiers on the balanced six-class OpenWhistle benchmark with session-disjoint train, validation, and test splits.",
+    "Validated use case: linear probing or finetuning of audio representation models for fine-grained intra-species dolphin whistle discrimination.",
+    "Supported exploratory use case: analysis of ten-class whistle categories using the unbalanced configuration, with explicit attention to class imbalance.",
+    "Not validated use case: inferring behavioral intent, communicative meaning, animal welfare state, individual identity beyond the released label definitions, or deployment performance in other sites without additional biological validation."
+  ],
+  "rai:dataLimitations": [
+    "The balanced configuration is designed for six-class whistle-type classification and should not be treated as a complete representation of all dolphin whistle categories or all individuals at the site.",
+    "The unbalanced configuration exposes ten classes but contains strong class imbalance, including rare whistle types that may not support reliable standalone evaluation.",
+    "The dataset comes from a semi-natural site with a stable pod of known dolphins, so results may not generalize to fully wild populations, other species, other recording devices, or different acoustic environments without validation.",
+    "Labels describe whistle categories, not full behavioral intent, communicative meaning, animal state, or ecological outcomes.",
+    "Session-disjoint splits reduce session leakage, but downstream users should still avoid mixing sessions or configurations in ways that invalidate the benchmark protocol."
+  ],
+  "rai:dataBiases": [
+    "The dataset reflects recordings from Dolphin Reef, Eilat and may overrepresent acoustic, social, and environmental conditions at that semi-natural site.",
+    "The naturally available whistle repertoire is imbalanced: some signature whistle and non-signature whistle categories are much more frequent than others.",
+    "The main balanced configuration intentionally subsamples classes to equalize six labels, which improves benchmark comparability but no longer reflects natural production frequencies.",
+    "Expert-refined labels depend on the available template contours, annotation pipeline, and annotator judgments, so ambiguous or low-quality whistles may be underrepresented.",
+    "Examples with usable F0 tracks or clearer spectrograms may be easier to analyze and may not fully represent all noisy, overlapping, or low-SNR acoustic conditions."
+  ],
+  "rai:personalSensitiveInformation": [
+    "No human subjects are involved in the dataset. Public columns contain animal audio clips, timing fields, whistle labels, F0 tracks, F0 spectrograms, and recording/session metadata.",
+    "The data consists of underwater animal acoustic recordings. Users should nevertheless treat incidental non-target sounds as possible and avoid attempts to infer people, vessels, precise locations, or sensitive field-site details from the audio.",
+    "Whistle labels may be associated with known dolphin signature-whistle categories, but the public dataset does not include sensitive human information."
+  ],
   "rai:dataSocialImpact": "The dataset supports open research on dolphin bioacoustics, animal communication, passive acoustic monitoring, and fine-grained acoustic representation learning. Positive impacts include reproducible evaluation for dolphin whistle-type classification and tools that may support non-invasive monitoring. Misuse risks include overinterpreting whistle-type predictions as behavioral meaning, identity, or conservation evidence without biological validation, or deploying models trained on this dataset in other sites without local evaluation. The data collection was passive and non-invasive, did not interfere with dolphin behavior, and involved no human subjects.",
   "rai:hasSyntheticData": false,
-  "rai:dataReleaseMaintenancePlan": "The dataset is hosted publicly on Hugging Face under dolphinteam/OpenWhistle-Classification-Finetuning and released under CC-BY 4.0. The repository includes deterministic review-sample configurations for reviewer inspection. Future OpenWhistle releases may expand labels, audio coverage, contextual metadata, or evaluation protocols.",
+  "rai:dataReleaseMaintenancePlan": "The dataset is hosted publicly on Hugging Face under dolphinteam/OpenWhistle-Classification-Finetuning and released under CC-BY 4.0. Future OpenWhistle releases may expand labels, audio coverage, contextual metadata, or evaluation protocols.",
   "prov:wasDerivedFrom": [
     {
       "@id": "urn:openwhistle:classification-source-recordings-and-annotations:neurips-2026",
@@ -1887,7 +825,7 @@
         "@id": "https://www.wikidata.org/wiki/Q5227332"
       },
       "prov:label": "Feature extraction and dataset packaging",
-      "sc:description": "Whistle clips, class labels, timing metadata, F0 contours, F0 confidence values, rendered F0 spectrograms, and clip-level snr_db estimates where available were packaged into Hugging Face parquet configurations. The balanced, unbalanced, and all configurations were split into train, validation, and test sets with session-disjoint assignment.",
+      "sc:description": "Whistle clips, class labels, timing metadata, F0 contours, F0 confidence values, rendered F0 spectrograms, and clip-level snr_db estimates where available were packaged into Hugging Face parquet configurations. The balanced and unbalanced configurations were split into train, validation, and test sets with session-disjoint assignment.",
       "prov:wasAttributedTo": [
         {
           "@type": "prov:SoftwareAgent",
@@ -1902,22 +840,6 @@
           "sc:description": "OpenWhistle research team responsible for preprocessing, curation, and release."
         }
       ]
-    },
-    {
-      "@type": "prov:Activity",
-      "prov:type": {
-        "@id": "https://www.wikidata.org/wiki/Q3306762"
-      },
-      "prov:label": "Review sample generation",
-      "sc:description": "Deterministic review-sample configurations were generated after final train/validation/test splits. The balanced review sample preserves equal rows per class within each split; unbalanced and all review samples preserve the source class distribution proportionally.",
-      "prov:wasAttributedTo": [
-        {
-          "@type": "prov:Agent",
-          "@id": "openwhistle_authors",
-          "prov:label": "OpenWhistle authors",
-          "sc:description": "OpenWhistle research team responsible for creating reviewer inspection subsets."
-        }
-      ]
     }
   ]
 }

unbalanced-review-sample/test-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e602d5de8e44114620a1f49ace137df65d6333e278d4a7a30914f4e03af2ea60
-size 29362615

unbalanced-review-sample/train-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:082ef68c17cbdc160c9d6aa025376f9390625172a842be1f94085dfa9c639884
-size 137120865

unbalanced-review-sample/validation-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4e9f949c7a77f347428e3175b52c0331c2f445dfc88f5bdc276f19fbdfa77bab
-size 29491865

unbalanced/test-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4bb6b45adf487a257363d71723735f70a293d78b241e3ceca7bc08573a135024
-size 214132531

all-review-sample/test-00000-of-00001.parquet → unbalanced/test-00000-of-00002.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72b3b5ea87422576487646e2ef393e311c84dab4b21811d547f0e7979f375f96
-size 29099100
+oid sha256:0ce120c0ec2bfeb7c849adfbba3a1ebcdda59de9662eda260b29eabd4aace673
+size 255775314

all-review-sample/train-00000-of-00001.parquet → unbalanced/test-00001-of-00002.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:74bf2f19a7de82dcdd30e321ddafd5d3a75f86b33907d27fb40cbf2f3de7046d
-size 139358235
+oid sha256:76ffa3013d3c0b637e67232429c29718f277e5336934e39a09f77d34725d1728
+size 258498197

unbalanced/train-00000-of-00003.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ed0d6d5db6d9b42d440f80b57389c4320267b6023d586d27364184659998182f
-size 329910654

all-review-sample/validation-00000-of-00001.parquet → unbalanced/train-00000-of-00006.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c774f3d57eace7fb38d4997034cb5428507cb03a580c37b1e822982582d65b3
-size 30489802
+oid sha256:d4ce9d45632bf51d8fddcaa36bfae186af830eb15346b10d6ecd5469a7e58c74
+size 396410956

unbalanced/train-00001-of-00003.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7663d9ed0e7c2b1878b1cb6a9adbfd973827e53fd2f08a852315462733f3a18e
-size 329852451

balanced-review-sample/test-00000-of-00001.parquet → unbalanced/train-00001-of-00006.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f03dc8961b7d59854f32b17bb2d92f663d88bef7144d1ef174041963ff1de37
-size 29338890
+oid sha256:6127b74806a93645e5832ce9369efacfdb44f7f84becf3ac91e1e4ab3430d189
+size 400806831

unbalanced/train-00002-of-00003.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6363842f87e20f1472b369cb2e58e6781c0c97043988e12a56f87081e04236e0
-size 331438605

unbalanced/train-00002-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e92e6848c945de1532123e37a550948499658e680872a6a73103fc69ce217e4
+size 399237115

unbalanced/train-00003-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ceb2765aea1ba06dbfda2f7f9db043dbca0a96caaac34ad8714810748ccfb9b
+size 402878854

unbalanced/train-00004-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca1c47afb1f8be74838e4d8568ba71691318a048e3368c300f0d6b8dab88198d
+size 400890457

unbalanced/train-00005-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58af6c6557a3b4c81eb89ca0351467fcffe5006d8fee84723e8961152939a530
+size 400027063

unbalanced/validation-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d3e76645b386eac428f1c05e40cc1499ce8753293d28f30c68cca2f0fa437084
-size 213894468

unbalanced/validation-00000-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b7f2632aef45803a2a64c2a303e6604b0be50c92972726e270640fa060fa1ab
+size 256223932

unbalanced/validation-00001-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5af42be02dea645d400e2577809ffc725e3951c4250dc997ea33f557fad3870
+size 253396819