Add human response baseline (rater-anonymized)

.gitattributes

@@ -0,0 +1,60 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.avro filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,158 @@
+---
+license: cc-by-nc-4.0
+pretty_name: Friend Bench
+tags:
+  - audio
+  - video
+  - social-relationship-recognition
+  - friend-bench
+configs:
+  - config_name: default
+    data_files:
+      - split: val
+        path: val/metadata.jsonl
+---
+
+<div align="center">
+
+<img src="https://huggingface.co/datasets/fluid-concepts/friend-bench/resolve/main/banner.png" alt="Friend Bench Banner" width="800px">
+
+**A benchmark for inferring relationship status from thin-slice observations of dyadic interaction**
+
+<table>
+<tr>
+<td align="center"><a href="https://huggingface.co/datasets/facebook/seamless-interaction">🌐 Built on Seamless Interaction</a></td>
+</tr>
+</table>
+
+</div>
+
+Human relationships leave subtle traces in how people talk, move, and react to each other. **Friend Bench** asks: from a short (~20s) clip of two people talking, can a model - or a human - tell whether they're strangers, friends, family, coworkers, or romantic partners?
+
+- 🎧 Multi-modal: audio, video, and transcript for every clip
+- 🏷️ Two tasks: 6-way relationship type, and binary familiar/stranger
+- ⏱️ Stratified by temporal position (early/late) within each interaction
+- 🔬 Built on [Seamless Interaction](https://huggingface.co/datasets/facebook/seamless-interaction)'s recordings
+
+## 🚀 Quick Start
+
+```python
+import pandas as pd
+from huggingface_hub import hf_hub_download, snapshot_download
+
+# Load the metadata index
+path = hf_hub_download(repo_id="fluid-concepts/friend-bench", repo_type="dataset",
+                        filename="val/metadata.jsonl")
+df = pd.read_json(path, lines=True)
+df[["sample_id", "relationship", "relationship_collapsed", "temporal_position"]].head()
+
+# Download every file for one sample
+snapshot_download(repo_id="fluid-concepts/friend-bench", repo_type="dataset",
+                   allow_patterns=["val/sample_000/*"], local_dir="friend_bench_sample")
+```
+
+## 🎬 Example Clips
+
+<table>
+<tr>
+<td align="center" width="50%">
+
+**Friends** (`sample_037`)
+
+<video controls src="https://huggingface.co/datasets/fluid-concepts/friend-bench/resolve/main/val/sample_037/video.mp4" width="100%"></video>
+
+</td>
+<td align="center" width="50%">
+
+**Strangers** (`sample_018`)
+
+<video controls src="https://huggingface.co/datasets/fluid-concepts/friend-bench/resolve/main/val/sample_018/video.mp4" width="100%"></video>
+
+</td>
+</tr>
+</table>
+
+## 📦 Dataset Structure
+
+```
+val/
+├── metadata.jsonl          # one row per sample: all fields below + media paths
+├── sample_NNN/
+│   ├── audio.wav           # mono, 16kHz mixed audio
+│   ├── video.mp4           # side-by-side video of both participants
+│   ├── transcript.txt      # turn-by-turn transcript
+│   └── metadata.json       # fields below
+└── human_responses.jsonl   # human-rater baseline judgments (see Fields below)
+```
+
+### Tasks
+
+| Task | Type | Labels |
+|---|---|---|
+| `relationship_type` | Multi-class | `stranger`, `friends`, `coworkers`, `family`, `romantic_partner`, `familiar_other` (`relationship_collapsed`) |
+| `familiar_stranger` | Binary | `familiar`, `stranger` (`relationship`) |
+
+### Fields
+
+| Field | Description |
+|---|---|
+| `sample_id` | e.g. `sample_000` |
+| `pair_key` | Dyad identifier (`vendor_id` + `session_id`) |
+| `file_ids` | Source Seamless Interaction participant file IDs |
+| `relationship` / `relationship_detail` / `relationship_collapsed` | Ground-truth relationship labels |
+| `temporal_position` | `early` / `late` position within the source interaction (this split has no `middle`) |
+| `clip_start` / `clip_end` / `clip_duration` | Clip bounds in seconds, within the source interaction |
+| `n_turns` / `transcript` | Turn count and transcript within the clip window |
+| `interaction_prompt_a` / `interaction_prompt_b` | The task prompt each participant was given |
+| `gender_a` / `gender_b` / `gender_config` | Participant genders, and whether they match |
+| `audio_quality` | Per-speaker loudness (LUFS) before/after normalization, faint-speaker flag |
+| `source` | Provenance within the generation pipeline |
+
+### Human response fields (`human_responses.jsonl`)
+
+| Field | Description |
+|---|---|
+| `modality` | Which modality condition the rater judged (`Audio`/`Video`/`Text`) |
+| `sample_id` | Joins to the sample's `metadata.json` |
+| `rater_id` | Anonymized rater identifier (`rater_001`, ...) |
+| `human_response` | The rater's guess: `familiar` or `stranger` |
+| `correct` | Whether `human_response` matched the sample's ground-truth `relationship` |
+| `rt_ms` | Reaction time in milliseconds |
+| `age` / `gender` / `english` | Rater demographics (bucketed) |
+| `tsis_total` / `tsis_mean` / `tsis1`-`tsis7` | Rater's Tromsø Social Intelligence Scale self-report |
+| `visual_body` / `visual_eye` / `visual_facial` | Visual-cue reliance ratings (Video only) |
+| `vocal_prosody` / `vocal_timing` / `vocal_coord` | Vocal-cue reliance ratings (Audio/Video only) |
+| `verbal_content` / `verbal_style` | Verbal-content-cue reliance ratings |
+| `inter_coord` / `inter_responsive` / `inter_rapport` / `inter_impression` | Interpersonal-cue reliance ratings |
+| `other_environ` | Environmental-cue reliance rating (Video only) |
+
+### Splits
+
+| Split | Samples | Description |
+|---|---:|---|
+| `val` | 96 | EO ice-breaker interactions across 3 vendor sites (`V00`/`V02`/`V03`, 32 each), balanced 48/48 on `familiar`/`stranger` and 48/48 on `early`/`late` |
+
+⚠️ A `test` split (holding out labels for leaderboard scoring) is planned but not yet released.
+
+## ⚠️ Known Limitations
+
+- **EO-only for now**: this split only contains ice-breaker (`EO`) interactions, not later-session (`IPC`) interactions.
+
+## 📄 License
+
+Friend Bench is derived from [Seamless Interaction](https://huggingface.co/datasets/facebook/seamless-interaction) and is released under the same license, **CC-BY-NC 4.0** (Creative Commons Attribution-NonCommercial 4.0 International):
+
+- **Share** - copy and redistribute in any medium or format
+- **Adapt** - remix, transform, and build upon the material
+
+Under the following terms:
+- **Attribution** - credit Seamless Interaction and this benchmark, with a link to the license
+- **NonCommercial** - no commercial use without explicit permission
+
+## 📑 Citation
+
+No citation available yet.
+
+## 🙏 Acknowledgments
+
+Built on the [Seamless Interaction Dataset](https://huggingface.co/datasets/facebook/seamless-interaction) from Meta FAIR.

val/sample_000/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ecee8cb6e62427f2ac651de56659e494f4c039884b0eca35e5ec383627d3e5e
+size 683428

val/sample_000/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_000",
+  "pair_key": "V00_S249",
+  "vendor_id": "V00",
+  "session_id": 249,
+  "relationship": "familiar",
+  "relationship_detail": "dating/spouse/romantic_partner",
+  "relationship_collapsed": "romantic_partner",
+  "temporal_position": "late",
+  "file_ids": [
+    "V00_S0249_I00000489_P0340",
+    "V00_S0249_I00000489_P0341"
+  ],
+  "clip_start": 326.05,
+  "clip_end": 347.39,
+  "clip_duration": 21.34,
+  "n_turns": 4,
+  "transcript": "[325.7s–331.6s] A: Not just a war event, a time in history, the 1982 Olympics, just anything.\n[326.1s–328.4s] B: Hmm.\n[333.5s–334.3s] A: Period in time.\n[334.7s–347.4s] B: I would go to probably the 70s, 60s times where there were just more intellectual talks amongst people of all variations of colors and things of that nature.",
+  "interaction_prompt_a": "Your partner will ask you an either-or question. Choose and answer and discuss it with them.",
+  "interaction_prompt_b": "Ask your partner: \"Would you rather go into the past and meet your ancestors or go into the future and meet your great-great-grandchildren?\" Discuss with your partner.",
+  "gender_a": "M",
+  "gender_b": "F",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -41.95,
+    "speaker_lufs_b": -38.21,
+    "speaker_lufs_post_a": -20.0,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 0.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_000/transcript.txt

@@ -0,0 +1,4 @@
+[325.7s–331.6s] A: Not just a war event, a time in history, the 1982 Olympics, just anything.
+[326.1s–328.4s] B: Hmm.
+[333.5s–334.3s] A: Period in time.
+[334.7s–347.4s] B: I would go to probably the 70s, 60s times where there were just more intellectual talks amongst people of all variations of colors and things of that nature.

val/sample_000/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4be0fa3b73fbd1a9c262e00beb1733c470da3dadbe8ad10b8680bf4f144f12a8
+size 1532916

val/sample_001/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68aa7c02ed0f64fa5265f59cafc312659056eab726fd145c09f5c837f070ef5e
+size 630862

val/sample_001/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_001",
+  "pair_key": "V00_S277",
+  "vendor_id": "V00",
+  "session_id": 277,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "late",
+  "file_ids": [
+    "V00_S0277_I00000385_P0252",
+    "V00_S0277_I00000385_P0374"
+  ],
+  "clip_start": 140.74,
+  "clip_end": 160.45,
+  "clip_duration": 19.71,
+  "n_turns": 12,
+  "transcript": "[140.7s–142.0s] A: Like, what is his day-to-day like?\n[142.2s–142.4s] B: Yeah.\n[142.4s–146.5s] A: Not nothing creepy or no weirdo behavior.\n[145.5s–145.7s] B: Right.\n[149.1s–149.8s] B: That's funny.\n[150.4s–150.8s] B: Yeah.\n[151.0s–151.5s] B: I don't know.\n[151.5s–153.9s] B: I still don't know who I was bi on.\n[153.1s–153.3s] A: You know?\n[154.7s–155.1s] B: Okay.\n[155.4s–155.6s] A: OK.\n[160.0s–160.5s] B: Okay.",
+  "interaction_prompt_a": "Your partner will ask you an either-or question. Choose and answer and discuss it with them.",
+  "interaction_prompt_b": "Would you rather have the ability to fly or be invisible? Discuss with your partner.",
+  "gender_a": "M",
+  "gender_b": "F",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -41.95,
+    "speaker_lufs_b": -41.43,
+    "speaker_lufs_post_a": -20.0,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 0.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_001/transcript.txt

@@ -0,0 +1,12 @@
+[140.7s–142.0s] A: Like, what is his day-to-day like?
+[142.2s–142.4s] B: Yeah.
+[142.4s–146.5s] A: Not nothing creepy or no weirdo behavior.
+[145.5s–145.7s] B: Right.
+[149.1s–149.8s] B: That's funny.
+[150.4s–150.8s] B: Yeah.
+[151.0s–151.5s] B: I don't know.
+[151.5s–153.9s] B: I still don't know who I was bi on.
+[153.1s–153.3s] A: You know?
+[154.7s–155.1s] B: Okay.
+[155.4s–155.6s] A: OK.
+[160.0s–160.5s] B: Okay.

val/sample_001/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b1348a6deeae8fff57d3276ee0f90ebd2eff21205bc3b4a38d3dcc8abc58b8b
+size 4173734

val/sample_002/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b0f90a584722886f2c884033911da05baa522d61cd57daaeaaf955b7ea2f936
+size 620622

val/sample_002/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_002",
+  "pair_key": "V00_S310",
+  "vendor_id": "V00",
+  "session_id": 310,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "early",
+  "file_ids": [
+    "V00_S0310_I00000385_P0042",
+    "V00_S0310_I00000385_P0417"
+  ],
+  "clip_start": 11.37,
+  "clip_end": 30.76,
+  "clip_duration": 19.39,
+  "n_turns": 4,
+  "transcript": "[11.4s–13.9s] A: Would you rather have the ability to fly or be invisible?\n[15.0s–15.3s] B: Fly.\n[15.3s–30.8s] B: Yeah, being invisible is like, kinda lonely, boring, like, can't go on dates with people easily because, I mean it would be a true testament to the personality overlooks thing.\n[16.6s–17.2s] A: Good choice.",
+  "interaction_prompt_a": "Your partner will ask you an either-or question. Choose and answer and discuss it with them.",
+  "interaction_prompt_b": "Would you rather have the ability to fly or be invisible? Discuss with your partner.",
+  "gender_a": "F",
+  "gender_b": "M",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -46.97,
+    "speaker_lufs_b": -32.78,
+    "speaker_lufs_post_a": -23.0,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 3.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_002/transcript.txt

@@ -0,0 +1,4 @@
+[11.4s–13.9s] A: Would you rather have the ability to fly or be invisible?
+[15.0s–15.3s] B: Fly.
+[15.3s–30.8s] B: Yeah, being invisible is like, kinda lonely, boring, like, can't go on dates with people easily because, I mean it would be a true testament to the personality overlooks thing.
+[16.6s–17.2s] A: Good choice.

val/sample_002/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f6452772a5f3f75e2ec1367a821d13446a095717c5c76f32d5a3a66e2475abf
+size 1901116

val/sample_003/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33a200c646f7871ddc013a705ffd5c8ac920ed3ab78c0d178c491b0f3c9b7db8
+size 600824

val/sample_003/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_003",
+  "pair_key": "V00_S498",
+  "vendor_id": "V00",
+  "session_id": 498,
+  "relationship": "familiar",
+  "relationship_detail": "dating/spouse/romantic_partner",
+  "relationship_collapsed": "romantic_partner",
+  "temporal_position": "early",
+  "file_ids": [
+    "V00_S0498_I00000477_P0635",
+    "V00_S0498_I00000477_P0636"
+  ],
+  "clip_start": 1.2,
+  "clip_end": 19.97,
+  "clip_duration": 18.77,
+  "n_turns": 6,
+  "transcript": "[1.2s–7.2s] A: Would you rather have to always speak in rhymes or sing everything you say?\n[1.3s–9.9s] B: Would you rather have to always speak in rhymes or sing?\n[14.4s–17.4s] B: I think I'd rather speak in rhymes.\n[18.1s–18.6s] A: Okay.\n[18.7s–20.0s] B: That would be pretty sublime.\n[18.9s–23.2s] A: Yeah, you would, you're killing it.",
+  "interaction_prompt_a": "Ask your partner: \"Would you rather always have to speak in rhyme or sing everything you say?\" Discuss it with them.",
+  "interaction_prompt_b": "Your partner will pose a question to you. Have fun answering it.",
+  "gender_a": "F",
+  "gender_b": "M",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -41.63,
+    "speaker_lufs_b": -49.81,
+    "speaker_lufs_post_a": -20.0,
+    "speaker_lufs_post_b": -25.8,
+    "residual_gap_db": 5.8,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_003/transcript.txt

@@ -0,0 +1,6 @@
+[1.2s–7.2s] A: Would you rather have to always speak in rhymes or sing everything you say?
+[1.3s–9.9s] B: Would you rather have to always speak in rhymes or sing?
+[14.4s–17.4s] B: I think I'd rather speak in rhymes.
+[18.1s–18.6s] A: Okay.
+[18.7s–20.0s] B: That would be pretty sublime.
+[18.9s–23.2s] A: Yeah, you would, you're killing it.

val/sample_003/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5678064157be975474970b31af01d61cab475c341f5a45449232ef3add8b187d
+size 2050348

val/sample_004/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69625cc381ad7e4ade8345626c13130740a51a0e02d8955e3322c058b269fded
+size 637688

val/sample_004/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_004",
+  "pair_key": "V00_S523",
+  "vendor_id": "V00",
+  "session_id": 523,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "late",
+  "file_ids": [
+    "V00_S0523_I00000477_P0666",
+    "V00_S0523_I00000477_P0667"
+  ],
+  "clip_start": 63.98,
+  "clip_end": 83.89,
+  "clip_duration": 19.91,
+  "n_turns": 12,
+  "transcript": "[64.0s–65.2s] A: Don't you like speaking in rhyme?\n[66.5s–67.4s] B: Sure, I'd love to.\n[68.2s–68.8s] B: I love rhyming.\n[69.4s–70.1s] A: Well, let's hear it.\n[74.2s–75.7s] B: It's hard to just come up with it.\n[74.6s–76.2s] A: It's not so easy.\n[75.8s–76.4s] B: No, it isn't.\n[76.2s–76.3s] A: Yeah.\n[76.9s–77.2s] B: It isn't.\n[79.5s–82.3s] B: I guess I have a lot of respect for the rappers that can do that.\n[82.8s–82.9s] A: Yeah.\n[83.6s–83.9s] B: Okay.",
+  "interaction_prompt_a": "Ask your partner: \"Would you rather always have to speak in rhyme or sing everything you say?\" Discuss it with them.",
+  "interaction_prompt_b": "Your partner will pose a question to you. Have fun answering it.",
+  "gender_a": "F",
+  "gender_b": "M",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -41.33,
+    "speaker_lufs_b": -43.18,
+    "speaker_lufs_post_a": -20.0,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 0.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_004/transcript.txt

@@ -0,0 +1,12 @@
+[64.0s–65.2s] A: Don't you like speaking in rhyme?
+[66.5s–67.4s] B: Sure, I'd love to.
+[68.2s–68.8s] B: I love rhyming.
+[69.4s–70.1s] A: Well, let's hear it.
+[74.2s–75.7s] B: It's hard to just come up with it.
+[74.6s–76.2s] A: It's not so easy.
+[75.8s–76.4s] B: No, it isn't.
+[76.2s–76.3s] A: Yeah.
+[76.9s–77.2s] B: It isn't.
+[79.5s–82.3s] B: I guess I have a lot of respect for the rappers that can do that.
+[82.8s–82.9s] A: Yeah.
+[83.6s–83.9s] B: Okay.

val/sample_004/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1efefdfd3f29577c07dcce36e6b1cf8471153714b58845a6121c1bf3306622b
+size 1670960

val/sample_005/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e2e7428797c389d47e3daa50f45bae19371e7f9cd615b0c051a6ccff8aefd10
+size 636324

val/sample_005/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_005",
+  "pair_key": "V00_S629",
+  "vendor_id": "V00",
+  "session_id": 629,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "late",
+  "file_ids": [
+    "V00_S0629_I00000371_P0752",
+    "V00_S0629_I00000371_P0785"
+  ],
+  "clip_start": 109.69,
+  "clip_end": 129.56,
+  "clip_duration": 19.87,
+  "n_turns": 11,
+  "transcript": "[108.7s–109.8s] A: No, no, no, no, no.\n[109.7s–114.9s] B: And I'm always between two people that might not necessarily be.\n[114.9s–119.0s] B: I don't know, but I get them all the time.\n[115.1s–116.8s] A: How can you sit in a middle seat?\n[119.1s–119.4s] A: No.\n[119.8s–122.9s] B: So because of that, just teleport me, please.\n[123.4s–123.8s] A: Okay.\n[123.7s–124.1s] B: Immediately.\n[125.4s–126.5s] B: Get me to my destination.\n[126.5s–126.5s] B: OK.\n[127.0s–129.6s] A: Okay, so, okay, we're teleporting, right?",
+  "interaction_prompt_a": "Ask your partner: \"Would you rather have the power to teleport or the ability to time travel?\"",
+  "interaction_prompt_b": "Your partner will pose a question to you. Have fun answering it.",
+  "gender_a": "F",
+  "gender_b": "M",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -43.94,
+    "speaker_lufs_b": -37.43,
+    "speaker_lufs_post_a": -20.0,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 0.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_005/transcript.txt

@@ -0,0 +1,11 @@
+[108.7s–109.8s] A: No, no, no, no, no.
+[109.7s–114.9s] B: And I'm always between two people that might not necessarily be.
+[114.9s–119.0s] B: I don't know, but I get them all the time.
+[115.1s–116.8s] A: How can you sit in a middle seat?
+[119.1s–119.4s] A: No.
+[119.8s–122.9s] B: So because of that, just teleport me, please.
+[123.4s–123.8s] A: Okay.
+[123.7s–124.1s] B: Immediately.
+[125.4s–126.5s] B: Get me to my destination.
+[126.5s–126.5s] B: OK.
+[127.0s–129.6s] A: Okay, so, okay, we're teleporting, right?

val/sample_005/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47e9a45307e9ef62938ddfb53338a0537d4fca5497bf3098bd6141a98c24de40
+size 2283907

val/sample_006/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b22149dc9798beab926585c662c01002fa171ba2e5985db8c90885c7e35afc9a
+size 685476

val/sample_006/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_006",
+  "pair_key": "V00_S866",
+  "vendor_id": "V00",
+  "session_id": 866,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "early",
+  "file_ids": [
+    "V00_S0866_I00000307_P0582",
+    "V00_S0866_I00000307_P0979"
+  ],
+  "clip_start": 0.0,
+  "clip_end": 21.4,
+  "clip_duration": 21.4,
+  "n_turns": 10,
+  "transcript": "[0.0s–0.6s] A: right.\n[1.1s–6.3s] A: Would you rather speak to the past or send messages to the future?\n[7.3s–8.1s] B: Oh, that's a great one.\n[8.5s–9.5s] B: Speak to the past.\n[10.9s–13.3s] A: Or send messages to your future?\n[13.3s–13.8s] B: the future.\n[13.3s–13.4s] A: Yeah.\n[13.8s–15.1s] B: I'm always\n[15.5s–17.8s] A: Or you could ask a question to the future if you want.\n[20.1s–21.4s] B: I would like to stay away from the future.",
+  "interaction_prompt_a": "Ask your partner: \"Would you rather have the ability to speak to the past or send messages to the future?\" Discuss it with them.",
+  "interaction_prompt_b": "Your partner will pose a question to you. Have fun answering it.",
+  "gender_a": "F",
+  "gender_b": "M",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -45.82,
+    "speaker_lufs_b": -48.84,
+    "speaker_lufs_post_a": -21.8,
+    "speaker_lufs_post_b": -24.8,
+    "residual_gap_db": 3.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_006/transcript.txt

@@ -0,0 +1,10 @@
+[0.0s–0.6s] A: right.
+[1.1s–6.3s] A: Would you rather speak to the past or send messages to the future?
+[7.3s–8.1s] B: Oh, that's a great one.
+[8.5s–9.5s] B: Speak to the past.
+[10.9s–13.3s] A: Or send messages to your future?
+[13.3s–13.8s] B: the future.
+[13.3s–13.4s] A: Yeah.
+[13.8s–15.1s] B: I'm always
+[15.5s–17.8s] A: Or you could ask a question to the future if you want.
+[20.1s–21.4s] B: I would like to stay away from the future.

val/sample_006/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcfcac1c9f6c83e9efaefc3284101cd1aea4e9ac9db3acf547a6af8bba265089
+size 3183800

val/sample_007/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fcabd7d9c51b140154de95942fcfaa0276efced0bdcba535c44f4d28854baba
+size 665678

val/sample_007/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_007",
+  "pair_key": "V00_S886",
+  "vendor_id": "V00",
+  "session_id": 886,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "early",
+  "file_ids": [
+    "V00_S0886_I00000477_P0687",
+    "V00_S0886_I00000477_P0995"
+  ],
+  "clip_start": 0.1,
+  "clip_end": 20.9,
+  "clip_duration": 20.8,
+  "n_turns": 7,
+  "transcript": "[0.1s–5.9s] A: rather sing every word that you say or speak in rhyme and why?\n[7.4s–9.7s] B: Um, I would rather sing.\n[10.9s–11.8s] A: Why would you rather sing?\n[12.5s–15.4s] B: Um, because I don't know how to rhyme.\n[15.4s–19.0s] B: I tried rapping for a while and I sucked.\n[17.7s–20.9s] A: Yeah.\n[20.5s–23.7s] B: So I would just sing and singing sounds better.",
+  "interaction_prompt_a": "Ask your partner: \"Would you rather always have to speak in rhyme or sing everything you say?\" Discuss it with them.",
+  "interaction_prompt_b": "Your partner will pose a question to you. Have fun answering it.",
+  "gender_a": "M",
+  "gender_b": "F",
+  "gender_config": "mixed",
+  "audio_quality": {
+    "speaker_lufs_a": -43.9,
+    "speaker_lufs_b": -35.69,
+    "speaker_lufs_post_a": -20.0,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 0.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_007/transcript.txt

@@ -0,0 +1,7 @@
+[0.1s–5.9s] A: rather sing every word that you say or speak in rhyme and why?
+[7.4s–9.7s] B: Um, I would rather sing.
+[10.9s–11.8s] A: Why would you rather sing?
+[12.5s–15.4s] B: Um, because I don't know how to rhyme.
+[15.4s–19.0s] B: I tried rapping for a while and I sucked.
+[17.7s–20.9s] A: Yeah.
+[20.5s–23.7s] B: So I would just sing and singing sounds better.

val/sample_007/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a3a21a0d1c14a1cecca19574c9a5c2af521bf327b4901aaf1ff96a895dd8b55
+size 3834783

val/sample_008/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ee3c093c40fb5edb0f7baca3d2b9930e019bbe257cacb4068f14d3b114de099
+size 706638

val/sample_008/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_008",
+  "pair_key": "V00_S279",
+  "vendor_id": "V00",
+  "session_id": 279,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "late",
+  "file_ids": [
+    "V00_S0279_I00000385_P0119",
+    "V00_S0279_I00000385_P0376"
+  ],
+  "clip_start": 113.94,
+  "clip_end": 136.0,
+  "clip_duration": 22.06,
+  "n_turns": 8,
+  "transcript": "[113.9s–116.6s] B: So, like, if I run into a plane by accident, am I just done for?\n[118.4s–118.7s] B: I'm joking.\n[118.7s–120.6s] B: I assume I'm just still Isaiah, but I can fly.\n[121.1s–125.0s] A: Well, yes, I think you'll just be flying in pieces.\n[128.8s–129.1s] B: Okay.\n[129.6s–130.5s] A: Yeah, that works out.\n[129.7s–136.0s] B: Yeah.\n[130.5s–136.0s] A: Yeah.",
+  "interaction_prompt_a": "Your partner will ask you an either-or question. Choose and answer and discuss it with them.",
+  "interaction_prompt_b": "Would you rather have the ability to fly or be invisible? Discuss with your partner.",
+  "gender_a": "M",
+  "gender_b": "M",
+  "gender_config": "same",
+  "audio_quality": {
+    "speaker_lufs_a": -43.72,
+    "speaker_lufs_b": -36.65,
+    "speaker_lufs_post_a": -20.0,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 0.0,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_008/transcript.txt

@@ -0,0 +1,8 @@
+[113.9s–116.6s] B: So, like, if I run into a plane by accident, am I just done for?
+[118.4s–118.7s] B: I'm joking.
+[118.7s–120.6s] B: I assume I'm just still Isaiah, but I can fly.
+[121.1s–125.0s] A: Well, yes, I think you'll just be flying in pieces.
+[128.8s–129.1s] B: Okay.
+[129.6s–130.5s] A: Yeah, that works out.
+[129.7s–136.0s] B: Yeah.
+[130.5s–136.0s] A: Yeah.

val/sample_008/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73e68848e26aee6b830d0ee16873fee9dcd188e79b30b47431f966d05a5c9a5a
+size 3332343

val/sample_009/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5dfd57342d5d06035b3fbfeee4878cfe0643c8b0de45cd1e1b1be011bb9ff9f
+size 620622

val/sample_009/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_009",
+  "pair_key": "V00_S285",
+  "vendor_id": "V00",
+  "session_id": 285,
+  "relationship": "familiar",
+  "relationship_detail": "family-generic",
+  "relationship_collapsed": "family",
+  "temporal_position": "late",
+  "file_ids": [
+    "V00_S0285_I00000489_P0386",
+    "V00_S0285_I00000489_P0387"
+  ],
+  "clip_start": 136.75,
+  "clip_end": 156.13,
+  "clip_duration": 19.38,
+  "n_turns": 6,
+  "transcript": "[127.0s–140.0s] B: Yeah, I think that there would be like a lot of benefit to going back because we only really know up until like, yeah, and like we never met any of our great grandparents, I don't think.\n[136.8s–137.3s] A: And not even.\n[137.4s–141.2s] A: I don't think so.\n[142.8s–150.4s] B: And it probably would give more, like both would give more perspective, but I think the future one, like I think I would choose it because it'd be trippy, but.\n[149.9s–150.1s] A: Yeah.\n[155.8s–156.1s] B: Thank you.",
+  "interaction_prompt_a": "Your partner will ask you an either-or question. Choose and answer and discuss it with them.",
+  "interaction_prompt_b": "Ask your partner: \"Would you rather go into the past and meet your ancestors or go into the future and meet your great-great-grandchildren?\" Discuss with your partner.",
+  "gender_a": "F",
+  "gender_b": "F",
+  "gender_config": "same",
+  "audio_quality": {
+    "speaker_lufs_a": -45.58,
+    "speaker_lufs_b": -41.9,
+    "speaker_lufs_post_a": -21.6,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 1.6,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_009/transcript.txt

@@ -0,0 +1,6 @@
+[127.0s–140.0s] B: Yeah, I think that there would be like a lot of benefit to going back because we only really know up until like, yeah, and like we never met any of our great grandparents, I don't think.
+[136.8s–137.3s] A: And not even.
+[137.4s–141.2s] A: I don't think so.
+[142.8s–150.4s] B: And it probably would give more, like both would give more perspective, but I think the future one, like I think I would choose it because it'd be trippy, but.
+[149.9s–150.1s] A: Yeah.
+[155.8s–156.1s] B: Thank you.

val/sample_009/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76d841ca067889ac91c8c406284c6b1b8ec980aec1719cccad2e05f243fbd81b
+size 2429448

val/sample_010/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74b0784f70aab00742d2378f673c4eca0ae93bdf14984c2bb5da82015f5af21b
+size 619256

val/sample_010/metadata.json

@@ -0,0 +1,33 @@
+{
+  "sample_id": "sample_010",
+  "pair_key": "V00_S516",
+  "vendor_id": "V00",
+  "session_id": 516,
+  "relationship": "familiar",
+  "relationship_detail": "friends",
+  "relationship_collapsed": "friends",
+  "temporal_position": "late",
+  "file_ids": [
+    "V00_S0516_I00000371_P0027A",
+    "V00_S0516_I00000371_P0658A"
+  ],
+  "clip_start": 206.57,
+  "clip_end": 225.9,
+  "clip_duration": 19.33,
+  "n_turns": 14,
+  "transcript": "[206.6s–207.1s] A: Really?\n[207.1s–209.6s] B: Because last time I went, we didn't get to go that far north.\n[210.1s–210.5s] A: Okay.\n[210.8s–214.5s] B: I'd like to go there and way south.\n[214.6s–215.5s] B: I didn't go to Sicily.\n[215.7s–216.5s] B: I went to the middle.\n[216.7s–218.7s] A: Okay, what's in Lake Cuomo?\n[216.9s–218.0s] B: I'd love to see Sicily.\n[218.7s–218.9s] A: It is.\n[219.5s–221.1s] B: Just a beautiful frickin' lake.\n[221.6s–223.5s] B: Yeah, well, that's what I hear, but I've never been.\n[223.9s–224.4s] A: Okay.\n[224.5s–224.8s] B: But...\n[225.7s–225.9s] A: Okay.",
+  "interaction_prompt_a": "Ask your partner: \"Would you rather have the power to teleport or the ability to time travel?\"",
+  "interaction_prompt_b": "Your partner will pose a question to you. Have fun answering it.",
+  "gender_a": "F",
+  "gender_b": "F",
+  "gender_config": "same",
+  "audio_quality": {
+    "speaker_lufs_a": -44.47,
+    "speaker_lufs_b": -36.75,
+    "speaker_lufs_post_a": -20.5,
+    "speaker_lufs_post_b": -20.0,
+    "residual_gap_db": 0.5,
+    "faint_speaker": false
+  },
+  "source": "reused_from_eo_full"
+}

val/sample_010/transcript.txt

@@ -0,0 +1,14 @@
+[206.6s–207.1s] A: Really?
+[207.1s–209.6s] B: Because last time I went, we didn't get to go that far north.
+[210.1s–210.5s] A: Okay.
+[210.8s–214.5s] B: I'd like to go there and way south.
+[214.6s–215.5s] B: I didn't go to Sicily.
+[215.7s–216.5s] B: I went to the middle.
+[216.7s–218.7s] A: Okay, what's in Lake Cuomo?
+[216.9s–218.0s] B: I'd love to see Sicily.
+[218.7s–218.9s] A: It is.
+[219.5s–221.1s] B: Just a beautiful frickin' lake.
+[221.6s–223.5s] B: Yeah, well, that's what I hear, but I've never been.
+[223.9s–224.4s] A: Okay.
+[224.5s–224.8s] B: But...
+[225.7s–225.9s] A: Okay.

val/sample_010/video.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a35fc110f3bad1aee8ad500025ee3b999c5d4a4e5a76ac2cd8905560be98a818
+size 2727728

val/sample_011/audio.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35effb8ba59030cabface7b69d2cbf22210a03c9ceb565a1612eb38fd829be81
+size 643150