Data and README changes

CHANGELOG.md

@@ -2,6 +2,103 @@
 
 All notable changes to **`nvidia/PhysicalAI-VANTAGE-Bench`** on Hugging Face.
 
+## 2026-06-02
+- **Public benchmark ecosystem links added.** Documented the official
+  VANTAGE-Bench website, GitHub benchmark repository, and Hugging Face
+  leaderboard as the public entry points for benchmark overview, run
+  guides/submission formats, and ranked results.
+
+## 2026-05-28
+- **Added `scripts/run_lmudata.py`** — a participant-facing data-prep tool
+  that builds an inference-ready, no-ground-truth (no-GT) LMUData layout
+  compatible with VLMEvalKit. It covers all VANTAGE-Bench tasks (VQA, Event
+  Verification, DVC, Temporal, 2D Pointing, Astro2D, 2D Grounding/RefDrone,
+  and SOT), prepares data for VLMEvalKit `--mode infer` (no local scoring),
+  and supports both Hugging Face remote sourcing
+  (`nvidia/PhysicalAI-VANTAGE-Bench`) and direct local dataset-repo sourcing
+  (auto-detected when run from inside this repo, or via `--local-source`).
+  Includes SOT preparation (downloads SmartSpaces source videos and extracts
+  frames via `ffmpeg`) and RefDrone/Grounding image preparation.
+- **Added `scripts/RUN_LMUData.md`** — a participant onboarding guide covering
+  setup, the HF cache, disk-space requirements, copy-vs-symlink media modes,
+  per-task data preparation notes, SOT internals/prerequisites, and
+  troubleshooting.
+
+## 2026-05-27
+- **`data/pointing/` migrated to JSONL as the canonical annotation
+  format.** `data/pointing/VANTAGE_2DPointing.jsonl` (1,005 lines, one
+  sample object per line, 8 fields: `index, question_id, image_path,
+  question, A, B, C, D`) replaces `VANTAGE_2DPointing.tsv`, which has
+  been removed. The JSONL is a lossless conversion of the prior TSV —
+  same 1,005 samples, same field values, same row order, with `index`
+  widened from string to integer. No sample content changed and no
+  ground-truth fields (`answer`, `target_point`) were introduced. The
+  Hugging Face Dataset Viewer `pointing` config now resolves cleanly
+  alongside the other JSON/JSONL configs (the mixed `.tsv`/`.jsonl`
+  `configs:` block was triggering HF's JSON builder on the TSV and
+  failing with an `ArrowInvalid: JSON parse error`).
+- **`data/event_verification/data_jsons/annotations/*.json` unwrapped
+  to top-level lists.** Each of the four files
+  (`VANTAGE_EventVerification.json` — 67 entries,
+  `tailgating_location_a.json` — 28, `tailgating_location_b.json` — 22,
+  `warehouse_near_miss.json` — 46; 163 total) was rewritten from
+  `{"bcq": [...]}` to `[...]`, matching the
+  `vqa/data_jsons/annotations/` and
+  `temporal_localization/data_jsons/annotations/` layouts. Every sample
+  object (`{id, video, system_prompt, question}`) is preserved
+  byte-for-byte and entry order is unchanged. This lets the Hugging
+  Face Dataset Viewer row-expand the files to 163 (instead of
+  collapsing each top-level object to a single row, which produced
+  4 rows in the viewer).
+- **`data/pointing/VANTAGE_2DPointing.tsv` updated.** MCQ option
+  coordinates (columns `A`–`D`) are now expressed in the normalized
+  `0–1000` coordinate system (each cell is an `x,y` pair, with both
+  components in `[0, 1000]` relative to the image dimensions). The
+  public TSV ships only the question side: columns are
+  `index, question_id, image_path, question, A, B, C, D`; the
+  ground-truth fields (`answer`, `target_point`) are held server-side
+  and are not included in the released file.
+- **`data/event_verification/` flattened** to match the `vqa/` and
+  `temporal_localization/` layout. The `filtered/.../{metropolis,tailgating,warehouse_near_miss}`
+  subtree was removed; all 163 videos now live directly under
+  `data/event_verification/videos/`, and the four annotation JSONs were
+  moved + renamed to
+  `data/event_verification/data_jsons/annotations/{metropolis_event_verification,tailgating_location_a,tailgating_location_b,warehouse_near_miss}.json`.
+  Each `bcq[].video` is now the basename (e.g. `example.mp4`) and each
+  `bcq[].id` is the stem (e.g. `example`); all other fields, entry
+  order, and counts (163 total) are preserved. The `configs:` glob in
+  the top-level README is updated accordingly.
+- Renamed remaining active Metropolis-named annotation JSON artifacts to
+  VANTAGE naming for public dataset consistency:
+  `data/event_verification/data_jsons/annotations/metropolis_event_verification.json`
+  → `data/event_verification/data_jsons/annotations/VANTAGE_EventVerification.json`;
+  `data/vqa/data_jsons/annotations/Metropolis_VQA_Verification_Final_ITS_Data.json`
+  → `data/vqa/data_jsons/annotations/VANTAGE_VQA_Verification_Final_ITS_Data.json`.
+  File contents are unchanged; only filenames moved. The `configs:`
+  globs in the top-level README already match the new filenames.
+  `data/README.md` updated to reference the new event-verification
+  filename.
+- **`data/vqa/data_jsons/annotations/*.json` reduced to inference-oriented fields.**
+  Each entry now carries exactly `{q_uid, question, options}`. The
+  metadata fields `industry`, `event_type`, `start_time`, `end_time`,
+  `video_duration`, `task_type`, and `dimension` were removed across all
+  five files. Entry count (1,195) and values of the retained fields are
+  unchanged. Smoke-tested against VLMEvalKit's `VANTAGE_VQA` inference
+  preparation: TSV regeneration, prompt building, and video resolution
+  all pass.
+- **`data/temporal_localization/data_jsons/annotations/*.json` reduced
+  to inference-oriented fields.** Each entry now carries exactly
+  `{vid, question_id, question}` (key order preserved). The metadata
+  fields `industry`, `event_type`, `task_type`, and `duration` were
+  removed across all three files. Entry count (1,067) and values of the
+  retained fields are unchanged; the 1,067 `question_id`s remain
+  globally unique. Smoke-tested against VLMEvalKit's `VANTAGE_Temporal`
+  inference preparation: TSV regeneration, prompt building, and video
+  resolution all pass.
+- Event Verification annotations were left unchanged in this pass; the
+  current `data/event_verification/data_jsons/annotations/*.json`
+  schema is treated as already inference-appropriate.
+
 ## 2026-05-19
 
 - README YAML updated with a `configs:` block so the HF dataset viewer

README.md

@@ -2,6 +2,35 @@
 license: other
 license_name: nvidia-evaluation-data-license
 license_link: LICENSE
+language:
+- en
+pretty_name: VANTAGE-Bench
+size_categories:
+- 10K<n<100K
+task_categories:
+- visual-question-answering
+- video-text-to-text
+- image-text-to-text
+- object-detection
+- multiple-choice
+task_ids:
+- visual-question-answering
+- image-captioning
+- multiple-choice-qa
+tags:
+- video
+- image
+- text
+- multimodal
+- video-understanding
+- image-understanding
+- benchmark
+- evaluation
+- infrastructure-cameras
+- warehouse
+- smart-city
+- intelligent-transportation-systems
+- smart-spaces
 configs:
 - config_name: vqa
   data_files:
@@ -14,7 +43,7 @@ configs:
 - config_name: event_verification
   data_files:
   - split: test
-    path: data/event_verification/filtered/**/*.json
+    path: data/event_verification/data_jsons/annotations/*.json
 - config_name: referring
   data_files:
   - split: test
@@ -22,7 +51,7 @@ configs:
 - config_name: pointing
   data_files:
   - split: test
-    path: data/pointing/Vantage2DPointing.tsv
+    path: data/pointing/VANTAGE_2DPointing.jsonl
 - config_name: tracking
   data_files:
   - split: test
@@ -37,27 +66,99 @@ configs:
     path: data/dense_captioning/metadata.jsonl
 ---
 
-# VANTAGE-BENCH
+# VANTAGE-Bench
 
 *Video ANalysis Tasks Across Generalized Environments*
 
+**3 domains · 8 tasks · 35,027 annotations · 3,346 media samples · 42 GB**
+
+<img src="./assets/vantage_bench_tasks.png" alt="VANTAGE-Bench task overview across Semantic, Temporal, Spatial, and Spatio-Temporal understanding categories" width="100%">
+
 ## Dataset Description
 
-VANTAGE-BENCH is the first public benchmark purpose-built for evaluating visual understanding on video captured by fixed infrastructure cameras. It spans three real-world domains — warehouse, smart city / Intelligent Transportation Systems (ITS), and smart spaces — across six spatio-temporal video understanding tasks including video question answering (VQA), temporal grounding, dense video captioning, event verification, spatial grounding, and spatio-temporal tracking.
+VANTAGE-Bench is the first public benchmark purpose-built for evaluating visual understanding on video captured by fixed infrastructure cameras. It spans three real-world domains — warehouse, smart city / Intelligent Transportation Systems (ITS), and smart spaces — across 8 tasks spanning semantic, temporal, spatial, and spatio-temporal evaluation, including video question answering (VQA), temporal localization, dense video captioning (DVC), event verification, spatial pointing, referring expressions, and spatio-temporal tracking. Unlike ordinary web video, this footage comes from fixed, infrastructure-mounted viewpoints — persistent scenes under long-duration monitoring that demand reasoning over stationary warehouse, ITS, and smart-space environments.
 
-This dataset is for evaluation purposes only.
+> **Evaluation-only / test split.** Ground-truth answers are withheld and all scoring is performed server-side — this dataset repository does not provide local benchmark scoring.
 
-## Dataset Owner(s)
+### Directory Structure
 
-NVIDIA Corporation
+```text
+PhysicalAI-VANTAGE-Bench/
+├── data/
+│   ├── 2dbbox/                # 2D object localization
+│   ├── dense_captioning/      # Dense video captioning
+│   ├── event_verification/    # Event verification
+│   ├── pointing/              # 2D spatial pointing
+│   ├── referring/             # 2D referring expressions
+│   ├── temporal_localization/ # Temporal localization
+│   ├── tracking/              # Single object tracking
+│   └── vqa/                   # Video question answering
+│
+├── scripts/
+│   ├── run_lmudata.py         # Prepare benchmark datasets
+│   └── RUN_LMUData.md         # Setup and usage guide
+│
+└── README.md                  # Dataset documentation
+```
 
-## Dataset Creation Date
+## Get Started
 
-April 24, 2026
+This repository contains the official VANTAGE-Bench dataset and data schemas. For benchmark documentation, submissions, and leaderboard results, use the resources below:
 
-## License/Terms of Use
+- **[VANTAGE-Bench's official website](https://vantage-bench.org/)** — detailed overview of VANTAGE-Bench, the benchmark suite, and submission entry points.
+- **[VANTAGE-Bench GitHub repository](https://github.com/Clemson-Capstone/VANTAGE-Bench)** — run guides, inference workflows, submission formats, and benchmark tooling.
+- **[Hugging Face leaderboard](https://huggingface.co/spaces/clemson-computing/VANTAGE-Bench-Leaderboard)** — currently ranked and highlighted models, and accepted user-submission results.
 
-This dataset is released under the [NVIDIA Evaluation Data License](./LICENSE.md).
+## Quick Start
+
+This repository ships the **test-split media and question-side annotations**;
+ground-truth answers are withheld for server-side scoring. VANTAGE-Bench's evaluation 
+toolkit expects benchmark datasets to be organized using a standard directory structure 
+called LMUData. To build an inference-ready LMUData layout across every task:
+
+```bash
+python scripts/run_lmudata.py --all --lmu-root ~/LMUData
+```
+
+Run from a clone of this dataset repo, the script auto-uses the local `data/`
+folder; otherwise it downloads the public dataset from Hugging Face. Two tasks
+fetch external media during this step — 2D Referring Expressions
+(RefDrone / VisDrone images) and Single Object Tracking (PhysicalAI-SmartSpaces
+videos; needs `ffmpeg` and an HF token). See
+[scripts/RUN_LMUData.md](./scripts/RUN_LMUData.md) for full setup, disk
+requirements, per-task notes, and troubleshooting.
+
+### What the setup produces
+
+`run_lmudata.py` automates the inference-prep step end to end. It sources the
+public dataset (an auto-detected local `data/` clone, an explicit
+`--local-source`, or a Hugging Face snapshot), builds each task's
+index file (`*.tsv` / `annotations.json`), and places the media by symlink
+(default) or `--copy`. It writes **no** ground-truth fields — withheld answers
+are left empty — and is idempotent, so re-runs only fill in what is missing.
+
+Most tasks need nothing beyond the command above. Two have extra prerequisites,
+which the script handles automatically when they are met:
+
+- **Single Object Tracking** — downloads source videos from `nvidia/PhysicalAI-SmartSpaces` and extracts frames with `ffmpeg`; needs an HF token with read access to that (gated) dataset.
+- **2D Referring Expressions (grounding)** — downloads the RefDrone / VisDrone images over the network.
+
+Under `--all`, a task that cannot meet its prerequisites is skipped while the
+others continue. The result is a inference-ready layout under
+`<LMUData root>/datasets/`:
+
+```text
+LMUData/
+└── datasets/
+    ├── Astro2D/
+    ├── VANTAGE_2DGrounding/
+    ├── VANTAGE_2DPointing/
+    ├── VANTAGE_DVC/
+    ├── VANTAGE_EventVerification/
+    ├── VANTAGE_SOT/
+    ├── VANTAGE_Temporal/
+    └── VANTAGE_VQA/
+```
 
 ## Dataset Characterization
 
@@ -67,21 +168,6 @@ Hybrid: Human, Synthetic, Automated. Video data is sourced from vendor-provided
 **Labeling Method**<br>
 Hybrid: Human, Synthetic, Pseudolabeled. Annotations for VQA, dense video captions, and temporal localization are primarily human-authored. Spatial grounding labels (2D/3D bounding boxes, referring expressions) use a combination of human annotation and pseudolabeling pipelines (detection + SAM for spatial pointing). Event verification labels are human-curated. Annotations are held server-side for evaluation only.
 
-### Directory Structure
-
-```text
-VANTAGE-BENCH/
-├── vqa/                     # Video question answering
-├── dense_captioning/        # Dense video captioning
-├── temporal_localization/   # Temporal localization
-├── event_verification/      # Event verification
-├── 2dbbox/                  # 2D object localization
-├── referring/               # 2D referring expressions
-├── pointing/                # 2D spatial pointing
-├── tracking/                # Spatio-temporal tracking
-└── README.md                # Dataset documentation and submission instructions
-```
-
 ## Evaluation
 
 ### Tasks and Submission Formats
@@ -89,34 +175,89 @@ VANTAGE-BENCH/
 | Category | Task | Metric |
 |----------|------|--------|
 | Semantic | VQA | Accuracy |
-| Semantic | Event Verification | F1 Score |
+| Semantic | Event Verification | Macro F1 |
 | Temporal | Dense Video Captioning | SODA-c |
-| Temporal | Temporal Localization | mAP@tIoU |
-| Spatial | 2D Object Localization | F1@0.5 |
+| Temporal | Temporal Localization | mIoU |
+| Spatial | 2D Object Localization | F1@IoU=0.5 |
 | Spatial | 2D Referring Expressions | mIoU |
-| Spatial | 2D Spatial Pointing | Pointing Accuracy |
+| Spatial | 2D Spatial Pointing | Accuracy |
 | Spatio-Temporal | Single Object Tracking | AUC |
 
-Expected submission formats and the leaderboard will be published soon.
+See [Submission Format](#submission-format) for the expected prediction schema. Results are submitted through the [official website](https://vantage-bench.org/) and ranked on the [Hugging Face leaderboard](https://huggingface.co/spaces/clemson-computing/VANTAGE-Bench-Leaderboard).
 
 ### Metric Notes
 
 - **Accuracy**: Percentage of correct predictions.
 - **SODA-c**: Metric for dense video captioning quality across event coverage and language quality.
-- **mAP@tIoU**: Mean Average Precision measured over temporal IoU thresholds.
-- **F1 Score**: Harmonic mean of precision and recall.
-- **F1@0.5**: F1 score at an IoU threshold of 0.5.
-- **mIoU**: Mean Intersection over Union — average overlap between predicted and ground-truth bounding boxes.
-- **Pointing Accuracy**: Percentage of correctly selected target regions.
+- **Macro F1**: Unweighted mean of per-class F1 scores (harmonic mean of precision and recall).
+- **F1@IoU=0.5**: F1 score at an IoU threshold of 0.5.
+- **mIoU**: Mean Intersection over Union — average overlap between predicted and ground-truth bounding boxes (also used for temporal localization spans).
 - **AUC**: Area under the ROC curve, measuring the model's ability to distinguish correct detections or tracks from incorrect ones across varying confidence thresholds.
 
-### Evaluation Server
+### Generating Predictions
+
+This dataset repository provides data and schemas only; it does **not** score
+predictions. Ground-truth answers are withheld and scoring happens server-side.
+The end-to-end workflow is:
+
+1. Prepare an inference-ready LMUData layout (see [Quick Start](#quick-start)):
+
+   ```bash
+   python scripts/run_lmudata.py --all --lmu-root ~/LMUData
+   ```
+
+2. Run inference using VANTAGE-Bench's evaluation toolkit. Each run emits a
+   `*.submission.jsonl` of predictions.
+3. Submit the predictions through the flow documented on
+   [VANTAGE-Bench's official website](https://vantage-bench.org/) and in the
+   [GitHub run guides](https://github.com/Clemson-Capstone/VANTAGE-Bench).
+   Accepted submissions are scored server-side and ranked on the
+   [Hugging Face leaderboard](https://huggingface.co/spaces/clemson-computing/VANTAGE-Bench-Leaderboard).
+
+See [scripts/RUN_LMUData.md](./scripts/RUN_LMUData.md) for setup, disk
+requirements, troubleshooting, and task-specific notes.
+
+### Submission Format
+
+Each prediction is a single JSON record (one record per line in the submission
+JSONL):
+
+```json
+{
+  "id": "<task_specific_id>",
+  "task": "<task_name>",
+  "conversations": [
+    {
+      "from": "assistant",
+      "value": "<raw_model_prediction>"
+    }
+  ],
+  "metadata": {
+    "model": "<model_name>",
+    "extra": {}
+  }
+}
+```
 
-The [VANTAGE-Bench GitHub repository](https://github.com/anon-benchmark/VANTAGE-bench) provides a sample evaluation pipeline for generating model predictions. Predictions are submitted to the official leaderboard, which will go live by the end of May 2026.
+The authoritative, per-task submission specification lives in the
+[GitHub repository](https://github.com/Clemson-Capstone/VANTAGE-Bench) and
+[scripts/RUN_LMUData.md](./scripts/RUN_LMUData.md). Submit through the entry
+points on the [official website](https://vantage-bench.org/); this repository
+performs no scoring or ranking.
 
 ## Dataset Format
 
-Video (mp4) and Images (jpg).
+Video (mp4) and images (jpg). Only the **input side** of each annotation ships; ground-truth answers are withheld. For example, a VQA record carries just the question and options:
+
+```json
+{
+  "q_uid": "GX010071_Clip_4.mp4",
+  "question": "How many people can exit the door at once while walking?",
+  "options": ["4", "all", "3", "2"]
+}
+```
+
+Field names vary by task (see [data/README.md](./data/README.md)); no record includes an answer or ground-truth label.
 
 ## Dataset Quantification
 
@@ -135,28 +276,29 @@ Video (mp4) and Images (jpg).
 **Total Media Samples (across tasks, with overlaps):** 3,346
 **Total Data Storage:** 42 GB
 
-## Potential Known Risks
+## Disclaimers
+
+### Potential Known Risks
 
 - Ground truth annotations are not publicly released. All evaluation is performed server-side.
 - Some warehouse videos are concatenated clips from longer recording sessions.
 
-## Citations
+### Ethical Considerations
 
-```bibtex
-@inproceedings{Fujita2020SODA,
-  author    = {Soichiro Fujita and Tsutomu Hirao and Hidetaka Kamigaito and Manabu Okumura and Masaaki Nagata},
-  title     = {{SODA}: Story Oriented Dense Video Captioning Evaluation Framework},
-  booktitle = {Proc. ECCV},
-  year      = {2020}
-}
+NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications.  When downloaded or used in accordance with our terms of service, developers should work with their internal developer teams to ensure this dataset meets requirements for the relevant industry and use case and addresses unforeseen product misuse.
 
-@inproceedings{Fu2024BLINK,
-  author    = {Xingyu Fu and Yushi Hu and Bangzheng Li and Yu Feng and Haoyu Wang and Xudong Lin and Dan Roth and Noah A. Smith and Wei-Chiu Ma and Ranjay Krishna},
-  title     = {{BLINK}: Multimodal Large Language Models Can See but Not Perceive},
-  booktitle = {Proc. ECCV},
-  year      = {2024}
-}
+Please report quality, risk, security vulnerabilities or NVIDIA AI Concerns [here](https://app.intigriti.com/programs/nvidia/nvidiavdp/detail).
 
+### References
+
+- [VANTAGE-Bench's official website](https://vantage-bench.org/)
+- [VANTAGE-Bench GitHub repository](https://github.com/Clemson-Capstone/VANTAGE-Bench)
+- [Hugging Face leaderboard](https://huggingface.co/spaces/clemson-computing/VANTAGE-Bench-Leaderboard)
+- [Hugging Face dataset](https://huggingface.co/datasets/nvidia/PhysicalAI-VANTAGE-Bench)
+
+### Citations
+
+```bibtex
 @article{Sun2025RefDrone,
   author    = {Zhichao Sun and Yuda Zou and Xian Sun and Yingchao Feng and Wenhui Diao and Menglong Yan and Kun Fu},
   title     = {{RefDrone}: A Challenging Benchmark for Referring Expression Comprehension in Drone Scenes},
@@ -165,16 +307,17 @@ Video (mp4) and Images (jpg).
 }
 ```
 
-## References
+### License/Terms of Use
+
+This dataset is released under the [NVIDIA Evaluation Data License](./LICENSE.md).
 
-- HuggingFace dataset: [nvidia/PhysicalAI-VANTAGE-Bench](https://huggingface.co/datasets/nvidia/PhysicalAI-VANTAGE-Bench)
+## Dataset Owner(s)
 
-<img src="./assets/vantage_bench_tasks.png" alt="VANTAGE-BENCH task overview across Semantic, Temporal, Spatial, and Spatio-Temporal understanding categories" width="100%">
+NVIDIA Corporation
 
-## Ethical Considerations
+## Dataset Creation Date
 
-NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications.  When downloaded or used in accordance with our terms of service, developers should work with their internal developer teams to ensure this dataset meets requirements for the relevant industry and use case and addresses unforeseen product misuse.
-Please report quality, risk, security vulnerabilities or NVIDIA AI Concerns [here](https://app.intigriti.com/programs/nvidia/nvidiavdp/detail).
+April 24, 2026
 
 ## Changelog
 

data/README.md

@@ -15,12 +15,10 @@ data/
 │   ├── prompt.json
 │   └── *.mp4
 ├── event_verification/           # Binary event classification
-│   └── filtered/
-│       ├── metropolis_event_verification/{*.mp4, test_annotation.json}
-│       ├── tailgating/{location_a, location_b}/{*.mp4, test_annotation.json}
-│       └── warehouse_near_miss/{*.mp4, test_annotations.json}
+│   ├── *.mp4 (under videos/)
+│   └── data_jsons/annotations/*.json
 ├── pointing/                     # 2D spatial pointing
-│   └── Vantage2DPointing.tsv
+│   └── VANTAGE_2DPointing.jsonl
 ├── referring/                    # 2D referring expressions
 │   └── refdrone_test_llava.json
 ├── temporal_localization/        # Temporal grounding
@@ -45,18 +43,49 @@ Tasks without a per-entry `question` field carry a top-level
 > Describe the notable events in the provided video. Provide the result in json format with `mm:ss.ff` format for time depiction for each event. Use keywords `start`, `end` and `caption` in the json output.
 
 ### `vqa/` — Video Question Answering
-Per-entry questions in `vqa/data_jsons/annotations/*.json`. Each entry has `{q_uid, question, options, …}`; answer keys (`gt`, `gt_option`, `*_update_*`, etc.) are removed.
+Per-entry questions in `vqa/data_jsons/annotations/*.json`. Each entry
+carries exactly three fields, scoped to model inference:
+- `q_uid` — video/sample identifier; resolves against `vqa/videos/`
+- `question` — natural-language question text
+- `options` — list of MCQ answer choices used to build the prompt
+
+Ground-truth (`gt_option`, `answer`) and per-question metadata
+(`industry`, `event_type`, `task_type`, `dimension`, `start_time`,
+`end_time`, `video_duration`) are not included in the public
+annotations.
 
 ### `temporal_localization/` — Temporal Grounding
-Per-entry questions in `temporal_localization/data_jsons/annotations/*.json`. Each entry has `{vid, question_id, question, duration, …}`; the `answer` timestamps are removed.
+Per-entry questions in `temporal_localization/data_jsons/annotations/*.json`.
+Each entry carries exactly three fields, scoped to model inference:
+- `vid` — video identifier; resolves against `temporal_localization/`
+- `question_id` — stable annotation identifier (reproducibility key)
+- `question` — temporal-localization query
+
+Ground-truth timestamps and per-question metadata (`industry`,
+`event_type`, `task_type`, `duration`) are not included in the public
+annotations.
 
 ### `event_verification/` — Binary Event Verification
-All four annotation files share a unified schema:
-`{"bcq": [{id, video, system_prompt, question}, …]}`. The binary `answer`
-is removed.
+Per-entry questions in `event_verification/data_jsons/annotations/*.json`
+(four files: `VANTAGE_EventVerification.json` — 67 entries,
+`tailgating_location_a.json` — 28, `tailgating_location_b.json` — 22,
+`warehouse_near_miss.json` — 46; 163 total). Each file is a top-level
+list of sample objects with schema
+`[{id, video, system_prompt, question}, …]` — matching the
+`vqa/` and `temporal_localization/` annotation layout — where `video`
+is the basename (e.g. `example.mp4`) and `id` is the stem
+(e.g. `example`), resolving against `event_verification/videos/`. The
+binary `answer` is removed.
 
 ### `pointing/` — 2D Spatial Pointing
-`Vantage2DPointing.tsv` — tab-separated. Each row carries the question and multiple-choice options; the last two ground-truth columns are dropped.
+`VANTAGE_2DPointing.jsonl` — one JSON object per line, 1,005 lines,
+8 fields: `index, question_id, image_path, question, A, B, C, D`. Each
+line carries the question and four multiple-choice options (`A`–`D`);
+each option is an `x,y` pair (string `"x,y"`) in the **normalized
+`0–1000` coordinate system** (both components in `[0, 1000]` relative
+to the image dimensions). `index` is an integer in `[0, 1004]`.
+Ground-truth fields (`answer`, `target_point`) are held server-side and
+are not included in the public JSONL.
 
 ### `referring/` — 2D Referring Expressions
 `refdrone_test_llava.json` — list of LLaVA conversation entries. Only the `human` turn (the question) is retained; the `gpt` turn (predicted bboxes) and GT meta fields are removed.

data/event_verification/data_jsons/annotations/VANTAGE_EventVerification.json

@@ -0,0 +1,404 @@
+[
+  {
+    "id": "LUPZNgg5idk_13",
+    "video": "LUPZNgg5idk_13.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a collision occur between two or more vehicles?"
+  },
+  {
+    "id": "IpgfZf6Y2BE_14",
+    "video": "IpgfZf6Y2BE_14.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a collision occur between two or more vehicles?"
+  },
+  {
+    "id": "IpgfZf6Y2BE_15",
+    "video": "IpgfZf6Y2BE_15.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a collision occur between two or more vehicles?"
+  },
+  {
+    "id": "NOALQmAB4yE_16",
+    "video": "NOALQmAB4yE_16.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a vehicle collide with pedestrian?"
+  },
+  {
+    "id": "SEb7p5oszeM_17",
+    "video": "SEb7p5oszeM_17.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a vehicle collide with a cyclist?"
+  },
+  {
+    "id": "SEb7p5oszeM_18",
+    "video": "SEb7p5oszeM_18.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a vehicle collide with a pedestrian?"
+  },
+  {
+    "id": "MmsgbcpWn-k_19",
+    "video": "MmsgbcpWn-k_19.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a collision occur between two or more vehicles?"
+  },
+  {
+    "id": "MmsgbcpWn-k_20",
+    "video": "MmsgbcpWn-k_20.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a collision occur between two or more vehicles?"
+  },
+  {
+    "id": "MmsgbcpWn-k_21",
+    "video": "MmsgbcpWn-k_21.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a collision occur between two or more vehicles?"
+  },
+  {
+    "id": "NOALQmAB4yE_24",
+    "video": "NOALQmAB4yE_24.mp4",
+    "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is “likely” if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a collision occur between two or more vehicles?"
+  },
+  {
+    "id": "evs_134db13b21",
+    "video": "evs_134db13b21.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_99c1cd175d",
+    "video": "evs_99c1cd175d.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_8cc3cd0258",
+    "video": "evs_8cc3cd0258.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_bc929d97da",
+    "video": "evs_bc929d97da.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_d897e4ada3",
+    "video": "evs_d897e4ada3.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_c3e684b820",
+    "video": "evs_c3e684b820.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_17560f2666",
+    "video": "evs_17560f2666.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_0f0c53aa1c",
+    "video": "evs_0f0c53aa1c.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_405dd1e5f8",
+    "video": "evs_405dd1e5f8.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_8f5ae5b865",
+    "video": "evs_8f5ae5b865.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_50815b9c8c",
+    "video": "evs_50815b9c8c.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "tailgating_13",
+    "video": "tailgating_13.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_866549be90",
+    "video": "evs_866549be90.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_e5ccfbd6bd",
+    "video": "evs_e5ccfbd6bd.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_d2523c5c64",
+    "video": "evs_d2523c5c64.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_f717d6dd57",
+    "video": "evs_f717d6dd57.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_110cbe0aac",
+    "video": "evs_110cbe0aac.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is anyone fighting?"
+  },
+  {
+    "id": "evs_8e472b1db0",
+    "video": "evs_8e472b1db0.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is anyone fighting?"
+  },
+  {
+    "id": "evs_abf9d9bc50",
+    "video": "evs_abf9d9bc50.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_c950abf04f",
+    "video": "evs_c950abf04f.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "Security_3_22",
+    "video": "Security_3_22.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is anyone fighting?"
+  },
+  {
+    "id": "Security_2_23",
+    "video": "Security_2_23.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "Security_2_24",
+    "video": "Security_2_24.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "Security_2_25",
+    "video": "Security_2_25.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Entering is from left to right and exiting is from right to left, is anyone exiting with a cart full of equipment?"
+  },
+  {
+    "id": "Security_2_26",
+    "video": "Security_2_26.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Entering is from left to right and exiting is from right to left, is anyone entering with a cart full of equipment?"
+  },
+  {
+    "id": "GX010011_Clip_8_27",
+    "video": "GX010011_Clip_8_27.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is the hallway overcrowded?"
+  },
+  {
+    "id": "GX010011_Clip_9_28",
+    "video": "GX010011_Clip_9_28.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Does everyone scan a badge to enter the room?"
+  },
+  {
+    "id": "evs_f262e9ed6a",
+    "video": "evs_f262e9ed6a.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_48a0587066",
+    "video": "evs_48a0587066.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "evs_191151ccf4",
+    "video": "evs_191151ccf4.mp4",
+    "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did a person tailgate through the access gate without badging?"
+  },
+  {
+    "id": "Warehouse_240219_GoPro_7_GX070600_100_3_2",
+    "video": "Warehouse_240219_GoPro_7_GX070600_100_3_2.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are all workers wearing PPE (hardhats and safety vests)?"
+  },
+  {
+    "id": "Warehouse_240219_GoPro_7_GX010600_500_2_3",
+    "video": "Warehouse_240219_GoPro_7_GX010600_500_2_3.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is the path obstructed for the forklift to pass?"
+  },
+  {
+    "id": "Warehouse_240219_GoPro_7_GX010600_500_1_4",
+    "video": "Warehouse_240219_GoPro_7_GX010600_500_1_4.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are the boxes properly stacked on the pallet loaded on the forklift?"
+  },
+  {
+    "id": "Warehouse_240219_GoPro_7_GX010600_400_1_5",
+    "video": "Warehouse_240219_GoPro_7_GX010600_400_1_5.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is the path obstructed for the forklift to pass?"
+  },
+  {
+    "id": "warehouse_1_600_4_6",
+    "video": "warehouse_1_600_4_6.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Did anyone experience a fall or end up on the ground?"
+  },
+  {
+    "id": "warehouse_1_540_7",
+    "video": "warehouse_1_540_7.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is any person near or in close proximity to the box when it falls?"
+  },
+  {
+    "id": "warehouse_1_540_4_8",
+    "video": "warehouse_1_540_4_8.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is the operator or a person using a cell phone while working?"
+  },
+  {
+    "id": "warehouse_1_425_6_9",
+    "video": "warehouse_1_425_6_9.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is the operator or a person jumping from the ladder?"
+  },
+  {
+    "id": "concat_wh_52_0_0_10",
+    "video": "concat_wh_52_0_0_10.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Does any box fall off the robot?"
+  },
+  {
+    "id": "warehouse_1_425_4_11",
+    "video": "warehouse_1_425_4_11.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are the boxes properly stacked as the operator lifts?"
+  },
+  {
+    "id": "warehouse_1_120_12",
+    "video": "warehouse_1_120_12.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Does the operator throw any boxes?"
+  },
+  {
+    "id": "concat_wh_52_0_5_13",
+    "video": "concat_wh_52_0_5_13.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Does a box fall off the robot?"
+  },
+  {
+    "id": "concat_wh_52_300_0_14",
+    "video": "concat_wh_52_300_0_14.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Does anyone walk in front of the forklift?"
+  },
+  {
+    "id": "concat_wh_52_300_1_15",
+    "video": "concat_wh_52_300_1_15.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Does anyone walk in front of the forklift?"
+  },
+  {
+    "id": "concat_wh_52_300_2_16",
+    "video": "concat_wh_52_300_2_16.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is the path of the forklift clear?"
+  },
+  {
+    "id": "concat_wh_52_300_2_17",
+    "video": "concat_wh_52_300_2_17.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are the boxes properly stacked on the pallet that is loaded on the forklift?"
+  },
+  {
+    "id": "concat_wh_52_300_1_18",
+    "video": "concat_wh_52_300_1_18.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are all workers wearing PPE?"
+  },
+  {
+    "id": "concat_wh_52_300_3_19",
+    "video": "concat_wh_52_300_3_19.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are the boxes properly stacked on the pallet that is loaded on the forklift?"
+  },
+  {
+    "id": "concat_wh_52_1890_0_20",
+    "video": "concat_wh_52_1890_0_20.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are all workers wearing PPE?"
+  },
+  {
+    "id": "concat_wh_52_1890_4_21",
+    "video": "concat_wh_52_1890_4_21.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are any boxes crushed?"
+  },
+  {
+    "id": "concat_wh_52_1890_4_22",
+    "video": "concat_wh_52_1890_4_22.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Do any boxes get dropped?"
+  },
+  {
+    "id": "concat_wh_52_1890_5_23",
+    "video": "concat_wh_52_1890_5_23.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Are any boxes crushed?"
+  },
+  {
+    "id": "concat_wh_52_1890_5_24",
+    "video": "concat_wh_52_1890_5_24.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Do any boxes get dropped?"
+  },
+  {
+    "id": "concat_wh_52_1890_9_25",
+    "video": "concat_wh_52_1890_9_25.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is everyone wearing a hardhat and safety vest?"
+  },
+  {
+    "id": "concat_wh_52_2925_1_26",
+    "video": "concat_wh_52_2925_1_26.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Do any boxes get dropped?"
+  },
+  {
+    "id": "concat_wh_52_2925_27",
+    "video": "concat_wh_52_2925_27.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is anything blocking the path of the small yellow robot?"
+  },
+  {
+    "id": "concat_wh_52_2925_28",
+    "video": "concat_wh_52_2925_28.mp4",
+    "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
+    "question": "Is anything blocking the path of the forklift?"
+  }
+]

data/event_verification/data_jsons/annotations/tailgating_location_a.json

@@ -0,0 +1,170 @@
+[
+  {
+    "id": "10_15_2025_sp_8_35_08_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_15_2025_sp_8_35_08_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_20_2025_sp_4_13_08_sp_PM_sp__lp_UTC-08_00_rp_",
+    "video": "11_20_2025_sp_4_13_08_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_24_2025_sp_9_44_18_sp_AM_sp__lp_UTC-08_00_rp_",
+    "video": "11_24_2025_sp_9_44_18_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_20_2025_sp_1_01_55_sp_PM_sp__lp_UTC-08_00_rp_",
+    "video": "11_20_2025_sp_1_01_55_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_30_25_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_30_25_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_15_2025_sp_8_38_19_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_15_2025_sp_8_38_19_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_8_2025_sp_8_38_03_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_8_2025_sp_8_38_03_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_09_05_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_09_05_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_46_58_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_46_58_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_24_2025_sp_10_23_17_sp_AM_sp__lp_UTC-08_00_rp_",
+    "video": "11_24_2025_sp_10_23_17_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_9_2025_sp_8_48_30_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_9_2025_sp_8_48_30_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_8_2025_sp_8_43_54_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_8_2025_sp_8_43_54_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_21_2025_sp_11_56_03_sp_AM_sp__lp_UTC-08_00_rp_",
+    "video": "11_21_2025_sp_11_56_03_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_16_2025_sp_9_00_13_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_16_2025_sp_9_00_13_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_48_55_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_48_55_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_12_31_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_12_31_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_16_2025_sp_9_10_29_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_16_2025_sp_9_10_29_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_15_2025_sp_8_39_16_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_15_2025_sp_8_39_16_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_24_2025_sp_10_44_28_sp_AM_sp__lp_UTC-08_00_rp_",
+    "video": "11_24_2025_sp_10_44_28_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_24_2025_sp_1_28_55_sp_PM_sp__lp_UTC-08_00_rp_",
+    "video": "11_24_2025_sp_1_28_55_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_32_48_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_32_48_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_21_2025_sp_11_40_45_sp_AM_sp__lp_UTC-08_00_rp_",
+    "video": "11_21_2025_sp_11_40_45_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_15_08_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_15_08_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_25_2025_sp_6_22_57_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_25_2025_sp_6_22_57_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_21_2025_sp_11_55_17_sp_AM_sp__lp_UTC-08_00_rp_",
+    "video": "11_21_2025_sp_11_55_17_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "10_9_2025_sp_8_51_57_sp_PM_sp__lp_UTC-07_00_rp_",
+    "video": "10_9_2025_sp_8_51_57_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_21_2025_sp_11_18_34_sp_AM_sp__lp_UTC-08_00_rp_",
+    "video": "11_21_2025_sp_11_18_34_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "11_24_2025_sp_2_57_33_sp_PM_sp__lp_UTC-08_00_rp_",
+    "video": "11_24_2025_sp_2_57_33_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  }
+]

data/event_verification/data_jsons/annotations/tailgating_location_b.json

@@ -0,0 +1,134 @@
+[
+  {
+    "id": "evs_6a52f11dad",
+    "video": "evs_6a52f11dad.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_b561420691",
+    "video": "evs_b561420691.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_907fe737cf",
+    "video": "evs_907fe737cf.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_0ea91247d8",
+    "video": "evs_0ea91247d8.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_6ad1a891ad",
+    "video": "evs_6ad1a891ad.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_d0e459f682",
+    "video": "evs_d0e459f682.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_2e30648c0a",
+    "video": "evs_2e30648c0a.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_292daa255e",
+    "video": "evs_292daa255e.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_a9e180fff3",
+    "video": "evs_a9e180fff3.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_53f64ccbe8",
+    "video": "evs_53f64ccbe8.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_b982d3f339",
+    "video": "evs_b982d3f339.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_03018e0ecf",
+    "video": "evs_03018e0ecf.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_bf746e9608",
+    "video": "evs_bf746e9608.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_6e738337bc",
+    "video": "evs_6e738337bc.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_f979eb0318",
+    "video": "evs_f979eb0318.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_024ae78480",
+    "video": "evs_024ae78480.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_fa68a5a4f8",
+    "video": "evs_fa68a5a4f8.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_eed8192951",
+    "video": "evs_eed8192951.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_32231b0bd6",
+    "video": "evs_32231b0bd6.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_a713802c9d",
+    "video": "evs_a713802c9d.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_3f674e8c19",
+    "video": "evs_3f674e8c19.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  },
+  {
+    "id": "evs_6a4da56832",
+    "video": "evs_6a4da56832.mp4",
+    "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
+    "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
+  }
+]

data/event_verification/data_jsons/annotations/warehouse_near_miss.json

@@ -0,0 +1,278 @@
+[
+  {
+    "id": "scene_07_01_00-23-52_to_00-25-33_GoPro1_Fork_Lift_stopped_while_person_crossing_the_isle_08-22",
+    "video": "scene_07_01_00-23-52_to_00-25-33_GoPro1_Fork_Lift_stopped_while_person_crossing_the_isle_08-22.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_08_01_00-00-46_to_00-02-20_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_01-15",
+    "video": "scene_08_01_00-00-46_to_00-02-20_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_01-15.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_08_02_00-02-20_to_00-04-56_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_08-22",
+    "video": "scene_08_02_00-02-20_to_00-04-56_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_08-22.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_08_03_00-04-56_to_00-08-15_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_04-18",
+    "video": "scene_08_03_00-04-56_to_00-08-15_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_04-18.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_09_01_00-08-15_to_00-10-22_GoPro1_Fork_Lift_moving_while_person_on_the_phone_crossing_the_isle_06-20",
+    "video": "scene_09_01_00-08-15_to_00-10-22_GoPro1_Fork_Lift_moving_while_person_on_the_phone_crossing_the_isle_06-20.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_10_01_00-10-22_to_00-13-12_GoPro1_Fork_Lift_moving_while_person_crossing_the_isle_06-20",
+    "video": "scene_10_01_00-10-22_to_00-13-12_GoPro1_Fork_Lift_moving_while_person_crossing_the_isle_06-20.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_11_01_00-13-12_to_00-16-16_GoPro1_Fork_Lift_moving_while_multiple_people_in_the_scene_04-22",
+    "video": "scene_11_01_00-13-12_to_00-16-16_GoPro1_Fork_Lift_moving_while_multiple_people_in_the_scene_04-22.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_12_01_00-16-16_to_00-17-31_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_06-20",
+    "video": "scene_12_01_00-16-16_to_00-17-31_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_06-20.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_12_02_00-17-31_to_00-19-50_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_02-16",
+    "video": "scene_12_02_00-17-31_to_00-19-50_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_02-16.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_14_01_00-19-50_to_00-22-54_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_04-18",
+    "video": "scene_14_01_00-19-50_to_00-22-54_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_04-18.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_16_01_00-22-54_to_00-25-38_GoPro1_person_walking_in_front_of_fork_lift_04-48",
+    "video": "scene_16_01_00-22-54_to_00-25-38_GoPro1_person_walking_in_front_of_fork_lift_04-48.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_17_01_00-25-38_to_00-27-32_GoPro1_person_running_in_front_of_fork_lift_02-16",
+    "video": "scene_17_01_00-25-38_to_00-27-32_GoPro1_person_running_in_front_of_fork_lift_02-16.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_17_02_00-27-32_to_00-30-55_GoPro1_person_running_in_front_of_fork_lift_02-26",
+    "video": "scene_17_02_00-27-32_to_00-30-55_GoPro1_person_running_in_front_of_fork_lift_02-26.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_18_01_00-30-55_to_00-33-57_GoPro1_person_jumping_to_not_get_hit_by_the_forklift_00-20",
+    "video": "scene_18_01_00-30-55_to_00-33-57_GoPro1_person_jumping_to_not_get_hit_by_the_forklift_00-20.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_19_01_00-33-57_to_00-35-41_GoPro1_fork_lift_moving_backwards_person_cutting_in_front_of_the_fo_04-24",
+    "video": "scene_19_01_00-33-57_to_00-35-41_GoPro1_fork_lift_moving_backwards_person_cutting_in_front_of_the_fo_04-24.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_20_01_00-35-41_to_00-37-23_GoPro1_fork_lift_going_backwards_person_running_passed_04-22",
+    "video": "scene_20_01_00-35-41_to_00-37-23_GoPro1_fork_lift_going_backwards_person_running_passed_04-22.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_21_01_00-37-23_to_00-39-30_GoPro1_fork_lift_going_backwards_person_stops_06-24",
+    "video": "scene_21_01_00-37-23_to_00-39-30_GoPro1_fork_lift_going_backwards_person_stops_06-24.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_22_01_00-39-30_to_00-45-36_GoPro1_fork_lift_moving_person_hesitating_and_stepping_back_01-20",
+    "video": "scene_22_01_00-39-30_to_00-45-36_GoPro1_fork_lift_moving_person_hesitating_and_stepping_back_01-20.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_23_01_00-45-36_to_00-49-03_GoPro1_boxes_blocking_the_view_of_the_driver_and_person_crossing_06-26",
+    "video": "scene_23_01_00-45-36_to_00-49-03_GoPro1_boxes_blocking_the_view_of_the_driver_and_person_crossing_06-26.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_25_01_00-02-02_to_00-05-05_GoPro1_person_working_on_boxes_while_fork_lift_approaches_06-30",
+    "video": "scene_25_01_00-02-02_to_00-05-05_GoPro1_person_working_on_boxes_while_fork_lift_approaches_06-30.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_26_01_00-05-05_to_00-08-50_GoPro1_same_as_above_person_jumping_05-24",
+    "video": "scene_26_01_00-05-05_to_00-08-50_GoPro1_same_as_above_person_jumping_05-24.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_27_01_00-08-50_to_00-15-25_GoPro1_person_bending_down_fork_lift_moving_forward_10-30",
+    "video": "scene_27_01_00-08-50_to_00-15-25_GoPro1_person_bending_down_fork_lift_moving_forward_10-30.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T1_C2_CS_S13T1_1-12-11-59_chunk_5__event_005_5",
+    "video": "Scene_13_S13T1_C2_CS_S13T1_1-12-11-59_chunk_5__event_005_5.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T3_C5_AS_S13T3_01-18-12-10_chunk_2__event_001_1",
+    "video": "Scene_13_S13T3_C5_AS_S13T3_01-18-12-10_chunk_2__event_001_1.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_5__event_008_8",
+    "video": "Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_5__event_008_8.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_6__event_001_1",
+    "video": "Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_6__event_001_1.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T4_C5_AS_S13T4_00-58-12-14_chunk_5__event_004_4",
+    "video": "Scene_13_S13T4_C5_AS_S13T4_00-58-12-14_chunk_5__event_004_4.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_1__event_004_4",
+    "video": "Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_1__event_004_4.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_5__event_004_4",
+    "video": "Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_5__event_004_4.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_6__event_001_1",
+    "video": "Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_6__event_001_1.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T5_C2_AS_S13T5_0-51-11-39_chunk_5__event_004_4",
+    "video": "Scene_13_S13T5_C2_AS_S13T5_0-51-11-39_chunk_5__event_004_4.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T5_C3_AS_S13T5_0-53-11-38_chunk_5__event_004_4",
+    "video": "Scene_13_S13T5_C3_AS_S13T5_0-53-11-38_chunk_5__event_004_4.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T5_C4_AS_S13T5_0-53-11-39_chunk_5__event_004_4",
+    "video": "Scene_13_S13T5_C4_AS_S13T5_0-53-11-39_chunk_5__event_004_4.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_13_S13T5_C6_AS_S13T5_00-29-11-15_chunk_5__event_003_3",
+    "video": "Scene_13_S13T5_C6_AS_S13T5_00-29-11-15_chunk_5__event_003_3.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_4_S4T4_C2_CV_S4T4_00-54-10-48_chunk_5__event_001_1",
+    "video": "Scene_4_S4T4_C2_CV_S4T4_00-54-10-48_chunk_5__event_001_1.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_4_S4T4_C4_CS_S4T4_00-54-10-48_chunk_4__event_003_3",
+    "video": "Scene_4_S4T4_C4_CS_S4T4_00-54-10-48_chunk_4__event_003_3.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_1__event_004_4",
+    "video": "Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_1__event_004_4.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_5__event_002_2",
+    "video": "Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_5__event_002_2.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_01_01_00-00-00_to_00-07-34_GoPro1_Calibration_with_people_walking_around_06-26",
+    "video": "scene_01_01_00-00-00_to_00-07-34_GoPro1_Calibration_with_people_walking_around_06-26.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_02_01_00-07-34_to_00-11-38_GoPro1_Forklifts_being_moved_out_of_the_way_06-26",
+    "video": "scene_02_01_00-07-34_to_00-11-38_GoPro1_Forklifts_being_moved_out_of_the_way_06-26.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_04_01_00-11-38_to_00-19-46_GoPro1_Forklift_entering_the_aisle_no_pedestrians_around_06-26",
+    "video": "scene_04_01_00-11-38_to_00-19-46_GoPro1_Forklift_entering_the_aisle_no_pedestrians_around_06-26.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_05_01_00-19-46_to_00-21-03_GoPro1_Fork_Lift_crossing_people_crossing_afterwards_06-26",
+    "video": "scene_05_01_00-19-46_to_00-21-03_GoPro1_Fork_Lift_crossing_people_crossing_afterwards_06-26.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_06_01_00-21-03_to_00-23-52_GoPro1_Fork_Lift_crossing_people_following_the_forklift_06-26",
+    "video": "scene_06_01_00-21-03_to_00-23-52_GoPro1_Fork_Lift_crossing_people_following_the_forklift_06-26.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_28_01_00-15-25_to_00-27-58_GoPro1_boxes_falling_04-22",
+    "video": "scene_28_01_00-15-25_to_00-27-58_GoPro1_boxes_falling_04-22.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_29_01_00-01-07_to_00-10-18_GoPro1_driver_picks_up_trash_00-50",
+    "video": "scene_29_01_00-01-07_to_00-10-18_GoPro1_driver_picks_up_trash_00-50.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  },
+  {
+    "id": "scene_29_01_00-09-27_to_00-12-53_GoPro1_driver_picks_up_trash_00-20",
+    "video": "scene_29_01_00-09-27_to_00-12-53_GoPro1_driver_picks_up_trash_00-20.mp4",
+    "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
+    "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
+  }
+]

data/event_verification/filtered/metropolis_event_verification/test_annotation.json

@@ -1,406 +0,0 @@
-{
-  "bcq": [
-    {
-      "id": "traffic_chunks/LUPZNgg5idk_13",
-      "video": "traffic_chunks/LUPZNgg5idk_13.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a collision occur between two or more vehicles?"
-    },
-    {
-      "id": "traffic_chunks/IpgfZf6Y2BE_14",
-      "video": "traffic_chunks/IpgfZf6Y2BE_14.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a collision occur between two or more vehicles?"
-    },
-    {
-      "id": "traffic_chunks/IpgfZf6Y2BE_15",
-      "video": "traffic_chunks/IpgfZf6Y2BE_15.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a collision occur between two or more vehicles?"
-    },
-    {
-      "id": "traffic_chunks/NOALQmAB4yE_16",
-      "video": "traffic_chunks/NOALQmAB4yE_16.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a vehicle collide with pedestrian?"
-    },
-    {
-      "id": "traffic_chunks/SEb7p5oszeM_17",
-      "video": "traffic_chunks/SEb7p5oszeM_17.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a vehicle collide with a cyclist?"
-    },
-    {
-      "id": "traffic_chunks/SEb7p5oszeM_18",
-      "video": "traffic_chunks/SEb7p5oszeM_18.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a vehicle collide with a pedestrian?"
-    },
-    {
-      "id": "traffic_chunks/MmsgbcpWn-k_19",
-      "video": "traffic_chunks/MmsgbcpWn-k_19.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a collision occur between two or more vehicles?"
-    },
-    {
-      "id": "traffic_chunks/MmsgbcpWn-k_20",
-      "video": "traffic_chunks/MmsgbcpWn-k_20.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a collision occur between two or more vehicles?"
-    },
-    {
-      "id": "traffic_chunks/MmsgbcpWn-k_21",
-      "video": "traffic_chunks/MmsgbcpWn-k_21.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a collision occur between two or more vehicles?"
-    },
-    {
-      "id": "traffic_chunks/NOALQmAB4yE_24",
-      "video": "traffic_chunks/NOALQmAB4yE_24.mp4",
-      "system_prompt": "You are a traffic monitoring system analyzing video of a street. Determine if a collision between vehicles, or vehicle and pedestrian or vehicle and cyclist has likely occurred.\nThe clip may not show the exact moment of impact, so use post-event evidence such as:\n- Vehicles in contact or showing visible damage (dents, debris, smoke, broken parts).\n- Pedestrian or cyclist on the ground, struck, or showing clear signs of impact or distress.. \n- Abnormal positions (intersecting, facing opposite directions, one against the side/rear of another).\n- Stationary vehicles remaining in contact or stopped in unnatural alignment.\n- Behavior inconsistent with normal driving (sudden halt, failure to separate, blocked motion).\n- Other unusual signs (e.g., airbags, leaking fluids, shattered glass) can also support the conclusion.\nA collision is \u201clikely\u201d if two or more independent cues strongly indicate impact, even if the collision itself is not shown. If evidence is weak or ambiguous, do not assume a collision.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a collision occur between two or more vehicles?"
-    },
-    {
-      "id": "safety_chunks/evs_134db13b21",
-      "video": "safety_chunks/evs_134db13b21.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_99c1cd175d",
-      "video": "safety_chunks/evs_99c1cd175d.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_8cc3cd0258",
-      "video": "safety_chunks/evs_8cc3cd0258.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_bc929d97da",
-      "video": "safety_chunks/evs_bc929d97da.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_d897e4ada3",
-      "video": "safety_chunks/evs_d897e4ada3.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_c3e684b820",
-      "video": "safety_chunks/evs_c3e684b820.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_17560f2666",
-      "video": "safety_chunks/evs_17560f2666.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_0f0c53aa1c",
-      "video": "safety_chunks/evs_0f0c53aa1c.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_405dd1e5f8",
-      "video": "safety_chunks/evs_405dd1e5f8.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_8f5ae5b865",
-      "video": "safety_chunks/evs_8f5ae5b865.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_50815b9c8c",
-      "video": "safety_chunks/evs_50815b9c8c.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/tailgating_13",
-      "video": "safety_chunks/tailgating_13.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_866549be90",
-      "video": "safety_chunks/evs_866549be90.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_e5ccfbd6bd",
-      "video": "safety_chunks/evs_e5ccfbd6bd.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_d2523c5c64",
-      "video": "safety_chunks/evs_d2523c5c64.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_f717d6dd57",
-      "video": "safety_chunks/evs_f717d6dd57.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_110cbe0aac",
-      "video": "safety_chunks/evs_110cbe0aac.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is anyone fighting?"
-    },
-    {
-      "id": "safety_chunks/evs_8e472b1db0",
-      "video": "safety_chunks/evs_8e472b1db0.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is anyone fighting?"
-    },
-    {
-      "id": "safety_chunks/evs_abf9d9bc50",
-      "video": "safety_chunks/evs_abf9d9bc50.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_c950abf04f",
-      "video": "safety_chunks/evs_c950abf04f.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/Security_3_22",
-      "video": "safety_chunks/Security_3_22.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is anyone fighting?"
-    },
-    {
-      "id": "safety_chunks/Security_2_23",
-      "video": "safety_chunks/Security_2_23.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/Security_2_24",
-      "video": "safety_chunks/Security_2_24.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/Security_2_25",
-      "video": "safety_chunks/Security_2_25.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Entering is from left to right and exiting is from right to left, is anyone exiting with a cart full of equipment?"
-    },
-    {
-      "id": "safety_chunks/Security_2_26",
-      "video": "safety_chunks/Security_2_26.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Entering is from left to right and exiting is from right to left, is anyone entering with a cart full of equipment?"
-    },
-    {
-      "id": "safety_chunks/GX010011_Clip_8_27",
-      "video": "safety_chunks/GX010011_Clip_8_27.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is the hallway overcrowded?"
-    },
-    {
-      "id": "safety_chunks/GX010011_Clip_9_28",
-      "video": "safety_chunks/GX010011_Clip_9_28.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Does everyone scan a badge to enter the room?"
-    },
-    {
-      "id": "safety_chunks/evs_f262e9ed6a",
-      "video": "safety_chunks/evs_f262e9ed6a.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_48a0587066",
-      "video": "safety_chunks/evs_48a0587066.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "safety_chunks/evs_191151ccf4",
-      "video": "safety_chunks/evs_191151ccf4.mp4",
-      "system_prompt": "You are a security monitoring system analyzing video of a access gate and hallways. \n\nGate Monitoring:\nWhen monitoring access gate, people are required to badge, the gate unlocks, and they enter. Determine whether a tailgating/piggybacking event has likely occurred (i.e., one or more people enter on a single authorization without individually badging).\nThe video clip shows the badge reader. Infer from post-event evidence such as:\n- Single gate-open event while multiple people pass through before the gate closes.\n- A follower enters closely behind the badged person (minimal gap in time or distance) without stopping at the reader or making a clear badging gesture.\n- The gate is held open/propped, or gate-open duration is longer than typical for a single entrant.\n- Multiple people cross the threshold during one gate cycle (gate does not close between them).\n- The leader looks back/holds the door while the follower does not badge.\n\nHallway Monitoring: \n- At the hallways look for fights and overcrowding.\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did a person tailgate through the access gate without badging?"
-    },
-    {
-      "id": "warehouse_chunks/Warehouse_240219_GoPro_7_GX070600_100_3_2",
-      "video": "warehouse_chunks/Warehouse_240219_GoPro_7_GX070600_100_3_2.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are all workers wearing PPE (hardhats and safety vests)?"
-    },
-    {
-      "id": "warehouse_chunks/Warehouse_240219_GoPro_7_GX010600_500_2_3",
-      "video": "warehouse_chunks/Warehouse_240219_GoPro_7_GX010600_500_2_3.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is the path obstructed for the forklift to pass?"
-    },
-    {
-      "id": "warehouse_chunks/Warehouse_240219_GoPro_7_GX010600_500_1_4",
-      "video": "warehouse_chunks/Warehouse_240219_GoPro_7_GX010600_500_1_4.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are the boxes properly stacked on the pallet loaded on the forklift?"
-    },
-    {
-      "id": "warehouse_chunks/Warehouse_240219_GoPro_7_GX010600_400_1_5",
-      "video": "warehouse_chunks/Warehouse_240219_GoPro_7_GX010600_400_1_5.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is the path obstructed for the forklift to pass?"
-    },
-    {
-      "id": "warehouse_chunks/warehouse_1_600_4_6",
-      "video": "warehouse_chunks/warehouse_1_600_4_6.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Did anyone experience a fall or end up on the ground?"
-    },
-    {
-      "id": "warehouse_chunks/warehouse_1_540_7",
-      "video": "warehouse_chunks/warehouse_1_540_7.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is any person near or in close proximity to the box when it falls?"
-    },
-    {
-      "id": "warehouse_chunks/warehouse_1_540_4_8",
-      "video": "warehouse_chunks/warehouse_1_540_4_8.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is the operator or a person using a cell phone while working?"
-    },
-    {
-      "id": "warehouse_chunks/warehouse_1_425_6_9",
-      "video": "warehouse_chunks/warehouse_1_425_6_9.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is the operator or a person jumping from the ladder?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_0_0_10",
-      "video": "warehouse_chunks/concat_wh_52_0_0_10.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Does any box fall off the robot?"
-    },
-    {
-      "id": "warehouse_chunks/warehouse_1_425_4_11",
-      "video": "warehouse_chunks/warehouse_1_425_4_11.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are the boxes properly stacked as the operator lifts?"
-    },
-    {
-      "id": "warehouse_chunks/warehouse_1_120_12",
-      "video": "warehouse_chunks/warehouse_1_120_12.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Does the operator throw any boxes?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_0_5_13",
-      "video": "warehouse_chunks/concat_wh_52_0_5_13.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Does a box fall off the robot?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_300_0_14",
-      "video": "warehouse_chunks/concat_wh_52_300_0_14.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Does anyone walk in front of the forklift?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_300_1_15",
-      "video": "warehouse_chunks/concat_wh_52_300_1_15.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Does anyone walk in front of the forklift?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_300_2_16",
-      "video": "warehouse_chunks/concat_wh_52_300_2_16.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is the path of the forklift clear?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_300_2_17",
-      "video": "warehouse_chunks/concat_wh_52_300_2_17.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are the boxes properly stacked on the pallet that is loaded on the forklift?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_300_1_18",
-      "video": "warehouse_chunks/concat_wh_52_300_1_18.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are all workers wearing PPE?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_300_3_19",
-      "video": "warehouse_chunks/concat_wh_52_300_3_19.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are the boxes properly stacked on the pallet that is loaded on the forklift?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_1890_0_20",
-      "video": "warehouse_chunks/concat_wh_52_1890_0_20.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are all workers wearing PPE?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_1890_4_21",
-      "video": "warehouse_chunks/concat_wh_52_1890_4_21.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are any boxes crushed?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_1890_4_22",
-      "video": "warehouse_chunks/concat_wh_52_1890_4_22.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Do any boxes get dropped?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_1890_5_23",
-      "video": "warehouse_chunks/concat_wh_52_1890_5_23.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Are any boxes crushed?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_1890_5_24",
-      "video": "warehouse_chunks/concat_wh_52_1890_5_24.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Do any boxes get dropped?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_1890_9_25",
-      "video": "warehouse_chunks/concat_wh_52_1890_9_25.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is everyone wearing a hardhat and safety vest?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_2925_1_26",
-      "video": "warehouse_chunks/concat_wh_52_2925_1_26.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Do any boxes get dropped?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_2925_27",
-      "video": "warehouse_chunks/concat_wh_52_2925_27.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is anything blocking the path of the small yellow robot?"
-    },
-    {
-      "id": "warehouse_chunks/concat_wh_52_2925_28",
-      "video": "warehouse_chunks/concat_wh_52_2925_28.mp4",
-      "system_prompt": "You are a warehouse monitoring system analyzing video footage. Your task is to answer safety and compliance questions strictly with \"Yes\" or \"No\".\nThe clip may not show the entire event, so rely on visible evidence. Infer from post-event evidence: \n- PPE compliance (hardhats, safety vests, etc.).\n- Path clear or obstructed for forklifts or robots.\n- Boxes stacked properly on pallets or being lifted.\n- Boxes crushed, dropped, or falling off forklifts/robots/operators.\n- Operator behavior (falling, using cell phone, throwing boxes).\n- Human safety risks (walking in front of forklift, near falling boxes, jumping from ladders).\n\nConfirm \"Yes\" only when visual evidence is clear.\nOtherwise, answer \"No\".",
-      "question": "Is anything blocking the path of the forklift?"
-    }
-  ]
-}

data/event_verification/filtered/tailgating/location_a/test_annotation.json

@@ -1,172 +0,0 @@
-{
-  "bcq": [
-    {
-      "id": "videos/site_1/category_tailgate/10_15_2025_sp_8_35_08_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_15_2025_sp_8_35_08_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_20_2025_sp_4_13_08_sp_PM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_20_2025_sp_4_13_08_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_24_2025_sp_9_44_18_sp_AM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_24_2025_sp_9_44_18_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_20_2025_sp_1_01_55_sp_PM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_20_2025_sp_1_01_55_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_30_25_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_30_25_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_15_2025_sp_8_38_19_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_15_2025_sp_8_38_19_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_8_2025_sp_8_38_03_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_8_2025_sp_8_38_03_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_09_05_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_09_05_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_46_58_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_46_58_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_24_2025_sp_10_23_17_sp_AM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_24_2025_sp_10_23_17_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_9_2025_sp_8_48_30_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_9_2025_sp_8_48_30_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_8_2025_sp_8_43_54_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_8_2025_sp_8_43_54_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/11_21_2025_sp_11_56_03_sp_AM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_tailgate/11_21_2025_sp_11_56_03_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_16_2025_sp_9_00_13_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_16_2025_sp_9_00_13_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_48_55_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_48_55_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_12_31_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_12_31_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_16_2025_sp_9_10_29_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_16_2025_sp_9_10_29_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_15_2025_sp_8_39_16_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_15_2025_sp_8_39_16_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_24_2025_sp_10_44_28_sp_AM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_24_2025_sp_10_44_28_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_24_2025_sp_1_28_55_sp_PM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_24_2025_sp_1_28_55_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_32_48_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_32_48_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_21_2025_sp_11_40_45_sp_AM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_21_2025_sp_11_40_45_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_15_08_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_15_08_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_25_2025_sp_6_22_57_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_25_2025_sp_6_22_57_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/11_21_2025_sp_11_55_17_sp_AM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_tailgate/11_21_2025_sp_11_55_17_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_tailgate/10_9_2025_sp_8_51_57_sp_PM_sp__lp_UTC-07_00_rp_",
-      "video": "videos/site_1/category_tailgate/10_9_2025_sp_8_51_57_sp_PM_sp__lp_UTC-07_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_21_2025_sp_11_18_34_sp_AM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_21_2025_sp_11_18_34_sp_AM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/site_1/category_badge/11_24_2025_sp_2_57_33_sp_PM_sp__lp_UTC-08_00_rp_",
-      "video": "videos/site_1/category_badge/11_24_2025_sp_2_57_33_sp_PM_sp__lp_UTC-08_00_rp_.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    }
-  ]
-}

data/event_verification/filtered/tailgating/location_b/test_annotation.json

@@ -1,136 +0,0 @@
-{
-  "bcq": [
-    {
-      "id": "videos/redacted/evs_6a52f11dad",
-      "video": "videos/redacted/evs_6a52f11dad.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_b561420691",
-      "video": "videos/redacted/evs_b561420691.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_907fe737cf",
-      "video": "videos/redacted/evs_907fe737cf.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_0ea91247d8",
-      "video": "videos/redacted/evs_0ea91247d8.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_6ad1a891ad",
-      "video": "videos/redacted/evs_6ad1a891ad.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_d0e459f682",
-      "video": "videos/redacted/evs_d0e459f682.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_2e30648c0a",
-      "video": "videos/redacted/evs_2e30648c0a.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_292daa255e",
-      "video": "videos/redacted/evs_292daa255e.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_a9e180fff3",
-      "video": "videos/redacted/evs_a9e180fff3.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_53f64ccbe8",
-      "video": "videos/redacted/evs_53f64ccbe8.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_b982d3f339",
-      "video": "videos/redacted/evs_b982d3f339.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_03018e0ecf",
-      "video": "videos/redacted/evs_03018e0ecf.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_bf746e9608",
-      "video": "videos/redacted/evs_bf746e9608.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_6e738337bc",
-      "video": "videos/redacted/evs_6e738337bc.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_f979eb0318",
-      "video": "videos/redacted/evs_f979eb0318.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_024ae78480",
-      "video": "videos/redacted/evs_024ae78480.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_fa68a5a4f8",
-      "video": "videos/redacted/evs_fa68a5a4f8.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_eed8192951",
-      "video": "videos/redacted/evs_eed8192951.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_32231b0bd6",
-      "video": "videos/redacted/evs_32231b0bd6.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_a713802c9d",
-      "video": "videos/redacted/evs_a713802c9d.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_3f674e8c19",
-      "video": "videos/redacted/evs_3f674e8c19.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    },
-    {
-      "id": "videos/redacted/evs_6a4da56832",
-      "video": "videos/redacted/evs_6a4da56832.mp4",
-      "system_prompt": "You are an expert AI assistant for video analysis. Your task is to accurately classify whether a surveillance video depicts **normal access** or **tailgating** behavior, based on the strict definitions below.\n\n# Category Definitions\n\n- **Normal Access**:  \nA person (or group) enters a secure area **properly authenticated** (e.g., by using a badge, keycard, or biometric system). Only those with authorization enter. There is **no evidence of security breach** or abnormal entry.  \n\n- **Tailgating**:  \nOne or more individuals **enter a secure area without authorization** by following closely behind an authorized person who has legitimately gained access. The unauthorized individual does not use valid credentials but takes advantage of the door being open. This includes:  \n- Following immediately behind someone with a badge swipe.  \n- Entering as part of a group where only the first person authenticates.  \n- Entering without any visible authentication while leveraging someone else's access.",
-      "question": "Does the video depict tailgating behavior? Answer \"Yes\" or \"No\"."
-    }
-  ]
-}

data/event_verification/filtered/warehouse_near_miss/test_annotations.json

@@ -1,280 +0,0 @@
-{
-  "bcq": [
-    {
-      "id": "positive/scene_07_01_00-23-52_to_00-25-33_GoPro1_Fork_Lift_stopped_while_person_crossing_the_isle_08-22",
-      "video": "positive/scene_07_01_00-23-52_to_00-25-33_GoPro1_Fork_Lift_stopped_while_person_crossing_the_isle_08-22.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_08_01_00-00-46_to_00-02-20_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_01-15",
-      "video": "positive/scene_08_01_00-00-46_to_00-02-20_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_01-15.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_08_02_00-02-20_to_00-04-56_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_08-22",
-      "video": "positive/scene_08_02_00-02-20_to_00-04-56_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_08-22.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_08_03_00-04-56_to_00-08-15_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_04-18",
-      "video": "positive/scene_08_03_00-04-56_to_00-08-15_GoPro1_Fork_Lift_crossing_while_person_running_in_front_of_it_04-18.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_09_01_00-08-15_to_00-10-22_GoPro1_Fork_Lift_moving_while_person_on_the_phone_crossing_the_isle_06-20",
-      "video": "positive/scene_09_01_00-08-15_to_00-10-22_GoPro1_Fork_Lift_moving_while_person_on_the_phone_crossing_the_isle_06-20.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_10_01_00-10-22_to_00-13-12_GoPro1_Fork_Lift_moving_while_person_crossing_the_isle_06-20",
-      "video": "positive/scene_10_01_00-10-22_to_00-13-12_GoPro1_Fork_Lift_moving_while_person_crossing_the_isle_06-20.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_11_01_00-13-12_to_00-16-16_GoPro1_Fork_Lift_moving_while_multiple_people_in_the_scene_04-22",
-      "video": "positive/scene_11_01_00-13-12_to_00-16-16_GoPro1_Fork_Lift_moving_while_multiple_people_in_the_scene_04-22.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_12_01_00-16-16_to_00-17-31_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_06-20",
-      "video": "positive/scene_12_01_00-16-16_to_00-17-31_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_06-20.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_12_02_00-17-31_to_00-19-50_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_02-16",
-      "video": "positive/scene_12_02_00-17-31_to_00-19-50_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_02-16.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_14_01_00-19-50_to_00-22-54_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_04-18",
-      "video": "positive/scene_14_01_00-19-50_to_00-22-54_GoPro1_Fork_Lift_moving_while_person_walking_behind_the_forklift_04-18.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_16_01_00-22-54_to_00-25-38_GoPro1_person_walking_in_front_of_fork_lift_04-48",
-      "video": "positive/scene_16_01_00-22-54_to_00-25-38_GoPro1_person_walking_in_front_of_fork_lift_04-48.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_17_01_00-25-38_to_00-27-32_GoPro1_person_running_in_front_of_fork_lift_02-16",
-      "video": "positive/scene_17_01_00-25-38_to_00-27-32_GoPro1_person_running_in_front_of_fork_lift_02-16.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_17_02_00-27-32_to_00-30-55_GoPro1_person_running_in_front_of_fork_lift_02-26",
-      "video": "positive/scene_17_02_00-27-32_to_00-30-55_GoPro1_person_running_in_front_of_fork_lift_02-26.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_18_01_00-30-55_to_00-33-57_GoPro1_person_jumping_to_not_get_hit_by_the_forklift_00-20",
-      "video": "positive/scene_18_01_00-30-55_to_00-33-57_GoPro1_person_jumping_to_not_get_hit_by_the_forklift_00-20.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_19_01_00-33-57_to_00-35-41_GoPro1_fork_lift_moving_backwards_person_cutting_in_front_of_the_fo_04-24",
-      "video": "positive/scene_19_01_00-33-57_to_00-35-41_GoPro1_fork_lift_moving_backwards_person_cutting_in_front_of_the_fo_04-24.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_20_01_00-35-41_to_00-37-23_GoPro1_fork_lift_going_backwards_person_running_passed_04-22",
-      "video": "positive/scene_20_01_00-35-41_to_00-37-23_GoPro1_fork_lift_going_backwards_person_running_passed_04-22.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_21_01_00-37-23_to_00-39-30_GoPro1_fork_lift_going_backwards_person_stops_06-24",
-      "video": "positive/scene_21_01_00-37-23_to_00-39-30_GoPro1_fork_lift_going_backwards_person_stops_06-24.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_22_01_00-39-30_to_00-45-36_GoPro1_fork_lift_moving_person_hesitating_and_stepping_back_01-20",
-      "video": "positive/scene_22_01_00-39-30_to_00-45-36_GoPro1_fork_lift_moving_person_hesitating_and_stepping_back_01-20.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_23_01_00-45-36_to_00-49-03_GoPro1_boxes_blocking_the_view_of_the_driver_and_person_crossing_06-26",
-      "video": "positive/scene_23_01_00-45-36_to_00-49-03_GoPro1_boxes_blocking_the_view_of_the_driver_and_person_crossing_06-26.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_25_01_00-02-02_to_00-05-05_GoPro1_person_working_on_boxes_while_fork_lift_approaches_06-30",
-      "video": "positive/scene_25_01_00-02-02_to_00-05-05_GoPro1_person_working_on_boxes_while_fork_lift_approaches_06-30.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_26_01_00-05-05_to_00-08-50_GoPro1_same_as_above_person_jumping_05-24",
-      "video": "positive/scene_26_01_00-05-05_to_00-08-50_GoPro1_same_as_above_person_jumping_05-24.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "positive/scene_27_01_00-08-50_to_00-15-25_GoPro1_person_bending_down_fork_lift_moving_forward_10-30",
-      "video": "positive/scene_27_01_00-08-50_to_00-15-25_GoPro1_person_bending_down_fork_lift_moving_forward_10-30.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T1_C2_CS_S13T1_1-12-11-59_chunk_5__event_005_5",
-      "video": "negative/Scene_13_S13T1_C2_CS_S13T1_1-12-11-59_chunk_5__event_005_5.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T3_C5_AS_S13T3_01-18-12-10_chunk_2__event_001_1",
-      "video": "negative/Scene_13_S13T3_C5_AS_S13T3_01-18-12-10_chunk_2__event_001_1.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_5__event_008_8",
-      "video": "negative/Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_5__event_008_8.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_6__event_001_1",
-      "video": "negative/Scene_13_S13T4_C4_AS_S13T4_1-12-12-28_chunk_6__event_001_1.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T4_C5_AS_S13T4_00-58-12-14_chunk_5__event_004_4",
-      "video": "negative/Scene_13_S13T4_C5_AS_S13T4_00-58-12-14_chunk_5__event_004_4.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_1__event_004_4",
-      "video": "negative/Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_1__event_004_4.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_5__event_004_4",
-      "video": "negative/Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_5__event_004_4.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_6__event_001_1",
-      "video": "negative/Scene_13_S13T4_C6_AS_S13T4_00-43-11-59_chunk_6__event_001_1.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T5_C2_AS_S13T5_0-51-11-39_chunk_5__event_004_4",
-      "video": "negative/Scene_13_S13T5_C2_AS_S13T5_0-51-11-39_chunk_5__event_004_4.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T5_C3_AS_S13T5_0-53-11-38_chunk_5__event_004_4",
-      "video": "negative/Scene_13_S13T5_C3_AS_S13T5_0-53-11-38_chunk_5__event_004_4.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T5_C4_AS_S13T5_0-53-11-39_chunk_5__event_004_4",
-      "video": "negative/Scene_13_S13T5_C4_AS_S13T5_0-53-11-39_chunk_5__event_004_4.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_13_S13T5_C6_AS_S13T5_00-29-11-15_chunk_5__event_003_3",
-      "video": "negative/Scene_13_S13T5_C6_AS_S13T5_00-29-11-15_chunk_5__event_003_3.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_4_S4T4_C2_CV_S4T4_00-54-10-48_chunk_5__event_001_1",
-      "video": "negative/Scene_4_S4T4_C2_CV_S4T4_00-54-10-48_chunk_5__event_001_1.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_4_S4T4_C4_CS_S4T4_00-54-10-48_chunk_4__event_003_3",
-      "video": "negative/Scene_4_S4T4_C4_CS_S4T4_00-54-10-48_chunk_4__event_003_3.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_1__event_004_4",
-      "video": "negative/Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_1__event_004_4.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_5__event_002_2",
-      "video": "negative/Scene_4_S4T4_C6_CS_S4T4_00-54-10-48_chunk_5__event_002_2.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_01_01_00-00-00_to_00-07-34_GoPro1_Calibration_with_people_walking_around_06-26",
-      "video": "negative/scene_01_01_00-00-00_to_00-07-34_GoPro1_Calibration_with_people_walking_around_06-26.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_02_01_00-07-34_to_00-11-38_GoPro1_Forklifts_being_moved_out_of_the_way_06-26",
-      "video": "negative/scene_02_01_00-07-34_to_00-11-38_GoPro1_Forklifts_being_moved_out_of_the_way_06-26.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_04_01_00-11-38_to_00-19-46_GoPro1_Forklift_entering_the_aisle_no_pedestrians_around_06-26",
-      "video": "negative/scene_04_01_00-11-38_to_00-19-46_GoPro1_Forklift_entering_the_aisle_no_pedestrians_around_06-26.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_05_01_00-19-46_to_00-21-03_GoPro1_Fork_Lift_crossing_people_crossing_afterwards_06-26",
-      "video": "negative/scene_05_01_00-19-46_to_00-21-03_GoPro1_Fork_Lift_crossing_people_crossing_afterwards_06-26.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_06_01_00-21-03_to_00-23-52_GoPro1_Fork_Lift_crossing_people_following_the_forklift_06-26",
-      "video": "negative/scene_06_01_00-21-03_to_00-23-52_GoPro1_Fork_Lift_crossing_people_following_the_forklift_06-26.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_28_01_00-15-25_to_00-27-58_GoPro1_boxes_falling_04-22",
-      "video": "negative/scene_28_01_00-15-25_to_00-27-58_GoPro1_boxes_falling_04-22.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_29_01_00-01-07_to_00-10-18_GoPro1_driver_picks_up_trash_00-50",
-      "video": "negative/scene_29_01_00-01-07_to_00-10-18_GoPro1_driver_picks_up_trash_00-50.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    },
-    {
-      "id": "negative/scene_29_01_00-09-27_to_00-12-53_GoPro1_driver_picks_up_trash_00-20",
-      "video": "negative/scene_29_01_00-09-27_to_00-12-53_GoPro1_driver_picks_up_trash_00-20.mp4",
-      "system_prompt": "You are an industrial safety analyst reviewing warehouse video. Determine whether the clip depicts a near-miss collision between a person and a forklift (or other powered vehicle). Answer Yes or No.",
-      "question": "Please tell whether the video contains near-miss collision between person and forklift. Your final answer should be either Yes or No."
-    }
-  ]
-}