The EUREQA dataset
In EUREQA, every question is constructed through an implicit reasoning chain. The chain is constructed by parsing DBPedia. Each layer comprises three components: an entity, a fact about the entity, and a relation between the entity
and its counterpart from the next layer. The layers stack up to create chains with different depths of reasoning. We verbalize reasoning chains into natural sentences and anonymize the entity of each layer to create the question.
Questions can be solved layer by layer and each layer is guaranteed a unique answer. EUREQA is not a knowledge game: we adopt a knowledge filtering process that ensures that most LLMs have sufficient world knowledge to answer our questions.
EUREQA comprises a total of 2,991 questions of different reasoning depths and difficulties. The entities encompass a broad spectrum of topics, effectively reducing any potential bias arising from specific entity categories.
These data are great for analyzing the reasoning processes of LLMs
PerformanceHere we present the accuracy of ChatGPT, Gemini-Pro and GPT-4 on the hard set of EUREQA across different depths d of reasoning (number of layers in the questions). We evaluate two prompt strategies: direct zero-shot prompt and ICL with two examples. In general, with the entities recursively substituted by the descriptions of reasoning chaining layers, and therefore eliminating surface-level semantic cues, these models generate more incorrect answers. When the reasoning depth increases from one to five on hard questions, there is a notable decline in performance for all models. This finding underscores the significant impact that semantic shortcuts have on the accuracy of responses, and it also indicates that GPT-4 is considerably more capable of identifying and taking advantage of these shortcuts.
| depth | d=1 | d=2 | d=3 | d=4 | d=5 | |||||
| direct | icl | direct | icl | direct | icl | direct | icl | direct | icl | |
| ChatGPT | 22.3 | 53.3 | 7.0 | 40.0 | 5.0 | 39.2 | 3.7 | 39.3 | 7.2 | 39.0 |
| Gemini-Pro | 45.0 | 49.3 | 29.5 | 23.5 | 27.3 | 28.6 | 25.7 | 24.3 | 17.2 | 21.5 |
| GPT-4 | 60.3 | 76.0 | 50.0 | 63.7 | 51.3 | 61.7 | 52.7 | 63.7 | 46.9 | 61.9 |
And in 2025, that way is a 1080p open matte scan with cinema DTS audio.
Through the dedicated work of fans archiving, scanning, color-correcting, and syncing original audio elements from private film collections, definitive pieces of cinematic history are saved from fading away. For fans of Spielberg's masterpiece, this version offers the ultimate nostalgic trip—a flawless recreation of the terror, awe, and celluloid beauty that captivated the world over three decades ago.
Given the next term, this likely means the creator framed their 1.85:1 transfer inside a 16:9 container, but “SuperWide” might imply they preserved the theatrical matte perfectly (1.85:1) with no additional cropping. For this release, treat it as , taken from the actual projection aperture. And in 2025, that way is a 1080p
However, Steven Spielberg and his cinematographer, Dean Cundey, intended the film to be seen in theaters at a widescreen ratio of 1.85:1. To achieve this, a physical mask was placed in the theater projector that blocked (or "matted out") the top and bottom of the frame. The official home video releases (DVD, Blu-ray, 4K) replicate this by "hard-matte" cropping the image, removing the top and bottom entirely.
Features the powerful LFE (subwoofer) track that made the T-Rex roar legendary. 🛠️ Why This Version Exists Purists often seek out these releases because: Given the next term, this likely means the
This specific file name refers to a of Steven Spielberg’s 1993 classic, Jurassic Park .
+----------------------------------------+ | UNSEEN OPEN MATTE AREA | -> (Reveals boom mics, stage equipment) +========================================+ | | | 1.85:1 THEATRICAL FRAMING | -> (What you see in standard Blu-rays) | | +========================================+ | UNSEEN OPEN MATTE AREA | -> (Reveals wires, lower set details) +----------------------------------------+ To achieve this, a physical mask was placed
: This is the most crucial visual element. It reveals that the film print used contains a taller frame or alternative aspect ratio than what was commonly seen on standard home video releases, offering a unique "open matte" viewing experience.
: This marks the initial public release of this specific community-led sync and color-correction effort. The Open Matte Phenomenon
: The definitive first complete release of this specific scanning and color-grading preservation project. The Open Matte Phenomenon: Seeing the Unseen
Projects like "jurassic.park.1993.35mm.1080p.cinema.dts.superwide.open.matte.v1.0" highlight a growing movement among cinephiles to protect film history. When studios alter films for modern displays, the original theatrical experience is often lost.
This website is adapted from Nerfies, UniversalNER and LLaVA, licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. We thank the LLaMA team for giving us access to their models.
Usage and License Notices: The data abd code is intended and licensed for research use only. They are also restricted to uses that follow the license agreement of LLaMA, ChatGPT, and the original dataset used in the benchmark. The dataset is CC BY NC 4.0 (allowing only non-commercial use) and models trained using the dataset should not be used outside of research purposes.