Upload evals-outputs/GAUNTLET.md with huggingface_hub
Browse files- evals-outputs/GAUNTLET.md +292 -0
evals-outputs/GAUNTLET.md
ADDED
@@ -0,0 +1,292 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# gauntlet results
|
2 |
+
|
3 |
+
These are are this model's output results on my "summarization gauntlet". You can find more info about that [here on my dropbox for it](https://www.dropbox.com/sh/axu1xlscrrexy55/AADAm01-4Zs3POyHQrgbDAsda?dl=0) or at [this dataset](https://huggingface.co/datasets/pszemraj/summcomparer-gauntlet-v0p1).
|
4 |
+
|
5 |
+
- if you aren't familiar with it, one thing to note is some of the docs **purposefully** are "messy"/have spelling errors etc.
|
6 |
+
|
7 |
+
parameters
|
8 |
+
|
9 |
+
```json
|
10 |
+
{
|
11 |
+
"model_name_or_path": "pszemraj/long-t5-tglobal-base-16384-book-summary",
|
12 |
+
"use_cuda": true,
|
13 |
+
"token_batch_length": 16384,
|
14 |
+
"batch_stride": 16,
|
15 |
+
"max_length_ratio": 0.25,
|
16 |
+
"load_in_8bit": false,
|
17 |
+
"compile_model": true,
|
18 |
+
"optimum_onnx": false,
|
19 |
+
"device": "cuda",
|
20 |
+
"inference_params": {
|
21 |
+
"min_length": 8,
|
22 |
+
"max_length": 4096,
|
23 |
+
"no_repeat_ngram_size": 3,
|
24 |
+
"encoder_no_repeat_ngram_size": 4,
|
25 |
+
"repetition_penalty": 2.5,
|
26 |
+
"num_beams": 10,
|
27 |
+
"num_beam_groups": 1,
|
28 |
+
"length_penalty": 1.0,
|
29 |
+
"early_stopping": true,
|
30 |
+
"do_sample": false
|
31 |
+
},
|
32 |
+
"textsum_version": "0.2.0"
|
33 |
+
}
|
34 |
+
```
|
35 |
+
|
36 |
+
- Created: `2023-11-28T20:06:52.034619`
|
37 |
+
|
38 |
+
## ASR-whisper-rpunctuated_Noam Chomsky, Fundam_1669853561_0_part1_summary
|
39 |
+
|
40 |
+
the narrator explains some of the foundational issues that are important to understanding language. He discusses the concept of "structure" in the human mind and how it allows us to express everything we know about the world through our thoughts. He uses examples from the 19th century to illustrate his theory of structure. The first problem is to find the internal languages of people who speak typologically diverse languages. The second task is to figure out how the speaker chooses an expression from the infinite array of sentences. There are two tasks that need to be undertaken: 1) determine the inner languages of speakers; and 2) find ways to explain the externalization of these internal languages. All of this work has been done over the years, but no one really understands what they're trying to do. For example, there are two kinds of constraints: 1) learnability and 2) evolvevability. If something can be explained by a device that can meet these two conditions, then why doesn't it solve the Yellow Land challenge? We don't yet know whether this explanation is genuine or not.
|
41 |
+
|
42 |
+
---
|
43 |
+
|
44 |
+
Section Scores for ASR-whisper-rpunctuated_Noam Chomsky, Fundam_1669853561_0_part1_summary:
|
45 |
+
|
46 |
+
- -1.1903
|
47 |
+
|
48 |
+
---
|
49 |
+
|
50 |
+
## ASR-whisper-rpunctuated_Noam Chomsky, Fundam_1669853631_0_part2_summary
|
51 |
+
|
52 |
+
the narrator explains some of the problems that have arisen from the use of merge in the scientific community since the early 1900s. The problem is that there is nothing to learn and no way to evolve the algorithms that yield the result. Merge, however, meets these conditions because there is something to learn. There are also other problems, such as how to connect these two different elements into a single workspace. To solve these problems, he uses the concept of "capital merge," which describes two things: first, a workspace that exists and forms a new workspace; and second, an object that does not exist but can be merged to it. This definition of merge was originally intended to limit the amount of resources that could be used by a merge operation. However, the original definition has now been extended to include all kinds of sub-cases, including limiting access. If you do not accept it, you'll get legitimateoperations, which yield invalid conclusions. For instance, if someone decides to create a set Xp, Yp, and then adds one of these men to it, they will violate all conditions for merging. That is, they won't be allowed to do anything more than what they've already done. Assuming that everyone has proposed this solution, the next step is to determine which class of merge operations are legitimate. First, find out which subclasses of merge are actually legitimate. Next, define merge according to general considerations -- general conditions that anyone linguistic operation should meet. And this should be deduced from them along with third factor properties,minimal computation,minivatable resources. All of this means that we have organized data rather than chaotic data. Finally, we need to formulate merge so that everything is conforming to the original intentions.
|
53 |
+
the UM uses the concept of "Pair Merge" to explain how one can combine two different objects into a single object. He then discusses some of the more common problems that people face when trying to figure out how to solve them. One problem is that there are certain types of words that cannot be extracted from another type of word. Another problem is head movement, which has been a problem for many years. Pisa introduces a new approach to dealing with this problem called "pair merge." This involves taking an object from one part of the brain and transforming it into another.
|
54 |
+
|
55 |
+
---
|
56 |
+
|
57 |
+
Section Scores for ASR-whisper-rpunctuated_Noam Chomsky, Fundam_1669853631_0_part2_summary:
|
58 |
+
|
59 |
+
- -1.2429
|
60 |
+
|
61 |
+
- -0.7433
|
62 |
+
|
63 |
+
---
|
64 |
+
|
65 |
+
## ASRnlp_law_lecture_week_1_v_2_c_transcription_1_summary
|
66 |
+
|
67 |
+
In this course, students will learn how to analyze and interpret documents as they are presented in the real world. They will also be introduced to some of the most important applications of machine learning and natural language processing. The first assignment is to read a response essay and write a report on it. The second assignment is an end-of-course assignment where students must submit a written paper within a month's time. Students who have completed all of their required readings should sign up for a two-credit course project designed to advance their knowledge of social science and machine learning.
|
68 |
+
|
69 |
+
---
|
70 |
+
|
71 |
+
Section Scores for ASRnlp_law_lecture_week_1_v_2_c_transcription_1_summary:
|
72 |
+
|
73 |
+
- -0.831
|
74 |
+
|
75 |
+
---
|
76 |
+
|
77 |
+
## ASRnlp_law_lecture_week_2_v_2_c_transcription_2_summary
|
78 |
+
|
79 |
+
In this first part of the course, students will learn how to analyze and classify documents using a dictionary. They will also be introduced to some of the most important concepts in social science, such as "tocanization" and "word frequency waiting."
|
80 |
+
|
81 |
+
---
|
82 |
+
|
83 |
+
Section Scores for ASRnlp_law_lecture_week_2_v_2_c_transcription_2_summary:
|
84 |
+
|
85 |
+
- -0.8407
|
86 |
+
|
87 |
+
---
|
88 |
+
|
89 |
+
## ASRnlp_law_lecture_week_3_part_1_v_2_c_transcription_3_summary
|
90 |
+
|
91 |
+
Tommo introduces the concept of feature selection and explains how it can be used to extract information from large, unsupervised sets of text. He discusses some of the social science applications that have been used to predict whether or not people will pay back a loan. He uses examples from four papers that use topic-based clustering to predict what kinds of legislation are likely to pass in different states.
|
92 |
+
|
93 |
+
---
|
94 |
+
|
95 |
+
Section Scores for ASRnlp_law_lecture_week_3_part_1_v_2_c_transcription_3_summary:
|
96 |
+
|
97 |
+
- -0.7136
|
98 |
+
|
99 |
+
---
|
100 |
+
|
101 |
+
## Emie_dissertation_cleansed_summary
|
102 |
+
|
103 |
+
This dissertation begins with a discussion of two American and British film noirs -- Act of Violence in 1948, The Men in 50, and Teresa in51 -- that relate to the question of what happens to the protagonists when they are confronted with the "cataclymic catastrophe" of the second world war. Enley sets off on a flight across America after his fellow veteran Joe Parkson is captured in a German prisoner's camp. On the way, Enley tells an officer about his plan to escape through a tunnel Dug with other men from the platoon; Enley kills himself and parkson in crashing his automobile. Reed's novel The Man Between takes place in Berlin, where Ivo Kern has become a supporter of the German government. When Susanne returns to Germany, Ivo sends her across the border checkspoints so that she will not be misidentified.
|
104 |
+
The narrator tells us that Reed's main goal in the film is to blur the borders between Berlin and New York. He uses a boy who climbs off his bicycle to create signs that Ivo or Susanne will not be coming. When they reach the train station, however, the police check both tickets and identity papers. They cannot get on the train because they are too distanced from the physical world of the East Sector. On the way back to their apartment, they confront their love for each other through "physical correspondences." In the basement room where Ivo has been kidnapped, she picks up an anvil to defend herself against Ivo. At this point, Ivo realizes that he must cross the border between two checkpoints in order to escape death. Meanwhile, while Ivo runs after the laundry truck, the boy jumps out of the back and shoots him. After he falls, he collapses in the snow. This scene contrasts with the staged elements of the final act of the film.
|
105 |
+
|
106 |
+
---
|
107 |
+
|
108 |
+
Section Scores for Emie_dissertation_cleansed_summary:
|
109 |
+
|
110 |
+
- -0.9879
|
111 |
+
|
112 |
+
- -1.1285
|
113 |
+
|
114 |
+
---
|
115 |
+
|
116 |
+
## OCR_ML4HLecture02image__summary
|
117 |
+
|
118 |
+
Ezurich introduces a number of different approaches to image analysis, including segmentation, superpixels, Markov random fields, and convolutional networks. These approaches are particularly useful in medical image analysis because they can be used to extract features from large-scale images while reducing the size of the original image.
|
119 |
+
|
120 |
+
---
|
121 |
+
|
122 |
+
Section Scores for OCR_ML4HLecture02image__summary:
|
123 |
+
|
124 |
+
- -0.537
|
125 |
+
|
126 |
+
---
|
127 |
+
|
128 |
+
## OCR_ML4HLecture04RepresentationLearning.pptx__summary
|
129 |
+
|
130 |
+
Ezurich introduces a new approach to learning time-representation in health care. He uses a novel representation-learning framework to develop an efficient method for learning time series from unsupervised and supervised approaches. He shows how well the proposed method is able to predict patients' health state over time. The results of the experiments are presented at the end of the paper.
|
131 |
+
|
132 |
+
---
|
133 |
+
|
134 |
+
Section Scores for OCR_ML4HLecture04RepresentationLearning.pptx__summary:
|
135 |
+
|
136 |
+
- -0.8881
|
137 |
+
|
138 |
+
---
|
139 |
+
|
140 |
+
## OCR_ML4HLecture05-NLP.pptx__summary
|
141 |
+
|
142 |
+
Ezrich explains how to use machine learning to automatically learn medical texts. He uses a bag of words as a representation of the word "bag" in order to create a model that predicts whether a word will appear in a given sentence. The results are quite similar to those obtained with other approaches, such as word-mapping and word-representation.
|
143 |
+
|
144 |
+
---
|
145 |
+
|
146 |
+
Section Scores for OCR_ML4HLecture05-NLP.pptx__summary:
|
147 |
+
|
148 |
+
- -0.7787
|
149 |
+
|
150 |
+
---
|
151 |
+
|
152 |
+
## OCR_PAPER_Hong et al. - 2022 - CogVideo Large-scale Pretraining for Text-to-Video Generation via Transformers-annotated__summary
|
153 |
+
|
154 |
+
Wu demonstrates the use of large-scale Pretrained Transformers, such as Cogview, to train text and image generation at the same time. The results show that Cog'Vie is superior to all publicly available methods in terms of speed, accuracy, and realism. This paper also includes a human evaluation of the model, which shows that it performs better than other open-source approaches.
|
155 |
+
|
156 |
+
---
|
157 |
+
|
158 |
+
Section Scores for OCR_PAPER_Hong et al. - 2022 - CogVideo Large-scale Pretraining for Text-to-Video Generation via Transformers-annotated__summary:
|
159 |
+
|
160 |
+
- -0.8845
|
161 |
+
|
162 |
+
---
|
163 |
+
|
164 |
+
## OCR_PAPER_Kandpal, Nieto, Jin - 2022 - Music Enhancement via Image Translation and Vocoding-annotated__summary
|
165 |
+
|
166 |
+
Oriol Neito develops a music-improvement model that uses conditional images to map low-quality musical signals into high-quality waves. He compares the listeners' scores with those of other musicians in the music field and concludes that these measures are poorly correlated with human perception.
|
167 |
+
|
168 |
+
---
|
169 |
+
|
170 |
+
Section Scores for OCR_PAPER_Kandpal, Nieto, Jin - 2022 - Music Enhancement via Image Translation and Vocoding-annotated__summary:
|
171 |
+
|
172 |
+
- -0.8371
|
173 |
+
|
174 |
+
---
|
175 |
+
|
176 |
+
## OCR_PAPER_dall-e-2-annotated__summary
|
177 |
+
|
178 |
+
the narrator discusses two approaches for image generation: first, by training a diffusion-decoder that can invert the Cllip image encoder to produce images; and second, by training a text-conditioned image generation stack that can be used as a scaffold for learning how to generate images. The first step is to train an auto-regressive or diffusion model on a captioned image, while the second steps are to train a decoctr that produces an image embedded in a given image. To further improve sample quality, we reduce the size of the original image by using principal component analysis, which finds that the ranks of the image representation space are dramatically reduced when training with SAM. This also results in improved sample quality because it allows us to sample from the "conditional distribution P1Zly" by sampling from the actual conditional distribution Zly. Next, we construct a latent space where we can directly see what the decipherer is seeing. These latent spaces allow us to manipulate the image into different types of variations, such as shapes and orientation. Finally, we compare both the diffusion-predictor and the pre-training version of UnClIP to determine which model performs better in terms of aesthetic quality.
|
179 |
+
|
180 |
+
---
|
181 |
+
|
182 |
+
Section Scores for OCR_PAPER_dall-e-2-annotated__summary:
|
183 |
+
|
184 |
+
- -1.1629
|
185 |
+
|
186 |
+
---
|
187 |
+
|
188 |
+
## The Most Dangerous Game--Richard Connell_summary
|
189 |
+
|
190 |
+
The narrator tells us that the ship is called "Ship Trap Island" because of its large, dark island. When Rainsford asks if they've passed the island yet, he can't figure out what it is. He doesn't know whether it's an evil place or a bad one. They talk for a while and then go back to sleep. In the middle of the night, rainsford hears a gun fired three times. He jumps onto the rail and balances himself on the rail. His pipe knocks him from his mouth when he realizes he's gone too far. He swims in the direction of the fireflies. After a few minutes, a small animal shoots him at the sea. As he swims away, ten minutes later, another sound comes from the sea: a pistol shot. It sounds like a big animal has killed a huge animal--Rainsford thinks there must have been twenty-two bullets fired by the brute. But we're not going to believe it until we see it again. We don't want to give you any clues as to what's going on, so he sinks down into the jungle. He wakes up feeling hungry. He looks around and sees lights in a towering chateau. He opens the door and finds General Zaroff holding a revolver. Ivan is dressed in evening clothes and sits down to dinner with the general. There are many animals on the table, including elephants, bears, lions, snakes, and other larger animals. This is kind of a weird thing to say, but this is actually a very good host. Rainsfor drinks a glass of port and talks about how he invented a hunting game. The general says he didn't really want to hunt anymore since he couldn't find a suitable animal to hunt. So he bought the island and built a house on it. He uses the island as a hideout where he spends all his time hunting. Finally, the general decides that he needs a new type of animal. He buys the island, builds the house, and sets up a training school for young Cossacks who are used to the deck. He gives them food and a knife, and they go hunting. One of the boys lost his head last night, and the rest of the group had no idea how to get off the rocks. The next day, General Zaloff comes to visit. He wants to know why Riversford wasn't feeling well. He was worried about losing his head after shooting a cowardly hunter. He also worries about the fact that these sailors are dull and stupid. He offers to help him leave the island at once. At midnight, GeneralZaloff sends Ivan along with some food, clothing, knives, and weapons. A couple of days later, Generalzaloff walks through the jungle looking forrainsford. On the way, Rainslet stops near a dead tree and begins to dig a hole in the ground. He cuts stakes and covers the hole with wild grapevine. Just as he gets close to the General's camp, hounds come closer to him. He climbs over a tree, holds a sapling, and fastens his hunting knife with the blade pointed down the trail. He pulls the sapling back and runs for his life without being hit. He manages to cross the gap between the trees and reach the shore.
|
191 |
+
|
192 |
+
---
|
193 |
+
|
194 |
+
Section Scores for The Most Dangerous Game--Richard Connell_summary:
|
195 |
+
|
196 |
+
- -1.5746
|
197 |
+
|
198 |
+
---
|
199 |
+
|
200 |
+
## gpt_peter_testing_group_exemplars_summary
|
201 |
+
|
202 |
+
the narrator explains some of the most important messages that have been sent to the UM over the past few weeks. The UM's main goal is to communicate with the other members of the community in order to learn more about their lives and ideas.
|
203 |
+
|
204 |
+
---
|
205 |
+
|
206 |
+
Section Scores for gpt_peter_testing_group_exemplars_summary:
|
207 |
+
|
208 |
+
- -0.9104
|
209 |
+
|
210 |
+
---
|
211 |
+
|
212 |
+
## navy seals copy pasta_summary
|
213 |
+
|
214 |
+
The narrator tells the audience that he's been training as a naval seals sniper and has killed hundreds of enemy soldiers. He also boasts about how he can kill anyone anywhere in the world with his "bare hands".
|
215 |
+
|
216 |
+
---
|
217 |
+
|
218 |
+
Section Scores for navy seals copy pasta_summary:
|
219 |
+
|
220 |
+
- -0.6729
|
221 |
+
|
222 |
+
---
|
223 |
+
|
224 |
+
## script_findingnemo_summary
|
225 |
+
|
226 |
+
In this short scene, the action shifts to Pixar's schoolhouse in Sydney. The students are greeted by their teacher, Mr. Ray, who tells them that Nemo is not ready to go to school because of his fear of the sea. He warns him that he will have to get him back before another fish swims out into the open water. When they see a white boat, however, they realize that it has been taken away by one of the divers. They all try to find their son, but he is nowhere to be found.
|
227 |
+
In this short scene, the action shifts to the seashore where Nemo and his father are searching for their son. When they find him, he tells them that he is sorry he couldn't make it back to his father because he was so eager to get out of the ocean. He also tells Nigel that his father has been fighting all sorts of fish, including sharks, jellyfish, and turtles. The group realizes that they have only 48 hours until Darla arrives, and they decide to clean up the fish tank in order to prevent her from getting there.
|
228 |
+
|
229 |
+
---
|
230 |
+
|
231 |
+
Section Scores for script_findingnemo_summary:
|
232 |
+
|
233 |
+
- -0.9441
|
234 |
+
|
235 |
+
- -0.9168
|
236 |
+
|
237 |
+
---
|
238 |
+
|
239 |
+
## script_frozendisney_summary
|
240 |
+
|
241 |
+
The action shifts to the Kingdom of Arendelle, where Elsa sleeps in bed with her little sister Anna. When she wakes up, she asks if she wants to build an iceman, and young Anna is too excited to do so. They play a game of ice-casting on the icefloor, but when they try to get one block out of the frozen water, it fails. Sven and Young Kristoff pull their ice blocks onto a horsedrawn sled and head off into the northern lights. The King's kingdom is located in a "honest castle," built of wood and nestled within a deep ravine. At night, the girls sneak into a ballroom, where they play snowballs, building snowmen, and making snowmen. In the morning, the King finds an old Norse map, scrambling to find a page that has an ancient map for him. The Queen and the King ride their horses away from the kingdom, leaving behind icy ice. On the other side of the valley, there are rocks that reveal bright faces. A troll named Buldas enters and tells the King that his powers will only grow as he shows a figure of an adult who creates magical snowflakes. He instructs them to learn how to control the spikes caused by human figures in the North Lights display. The two men decide to lock the gates of the castle and reduce the staff so that no one will be able to come in and out. Meanwhile, Anna sits in her room playing dolls, giving up hope of ever seeing her again. She asks whether she can build a new snowman, but she does not know what to do. One day, she freezes her windowsill because she doesn't have gloves on her hands. But the King slips some leather gloves onto her hand, which makes her feel better. The scene shifts back to Anne's room, where she sits at her window looking out through the locked door. Her room is frozen, and she cannot see anyone outside.
|
242 |
+
The narrator tells us that it's "true love". A wolf throws a blanket at the wolves, and then another wolf knocks him away. Sven is about to jump into a huge gorge when he sees an axe flying at him. Anna grabs her sled, but she doesn't realize what's happening. She says she'll replace everything in it if she dies. As they walk out of the forest, a snowman named Olaf shows up. He has a noseless snowman on his back who looks like he wants to be kissed by Anna. It turns out there's some kind of weird relationship going on between Olaf and Anna. Oh, yeah--olaf takes a carrot from one of their satches as he runs toward Anna. They have a long conversation about how Elsa needs to bring summer back to the kingdom. Everyone agrees that this is the best way for them to get back to summer. In the middle of this conversation, Prince Hans enters and asks if anyone needs a new cloak. When he learns that Arendelle will be giving away all of her goods to the Duke, he orders two men to go with him to find Anna. On the morning of the first day of the winter, Anna climbs onto the top of a steep mountain. Her legs are shaken, and she can't stand the fact that she's being chased down by a horse. Finally, she manages to escape through a stairway, which leads to a staircase leading to the main living quarters of the castle. Once inside, Anna finds herself standing on a balcony, looking at the beauty of the place. She notices that she looks different than she did when she was a young girl. She knows that she can do whatever she wants without hurting anyone. So she decides to build a giant snowman called Marshmallow so that she won't need to worry about keeping her distance during the winter. Meanwhile, Anna starts throwing snowballs at Marshallmallow, whom she thinks is stronger than she realizes. Just as everyone is getting ready to leave, Marmillallow pulls both Anna and Kristoff off the side of the hill. Konstantiniously, Krostoff and Anna fall right over the wall of the Castle. At this point, Kristoff gets worried about Anna's hair, which seems to be turning white. But not before he convinces her to stop worrying about his ice business. This time, instead of worrying about whether or not she should wear a cape or hat, she insists on running across the ice to meet Sven and his family. Soon after, however, Bulda comes up with a plan to make Anna look like a cowboy. Instead of trying to change her mind, she suggests that she would like to keep her heart full of ice, even if it were only for a short time. That way, she wouldn't freeze forever. If she didn't remove her heart, solid ice would freeze forever; otherwise, she'd freeze permanently. All of this is interrupted by the arrival of Prince Hans, who announces that he plans to return to Dover to search for Anna.
|
243 |
+
The scene shifts to the castle, where Olaf and Anna are trapped by a fierce winter storm. Sven attempts to lift him off the ground, but he does not succeed. A violent winter storm overtakes the castle; sharp ice cringes at the castle's ceiling. On the night of the storm, Anna is shivering by the door when she sees a carrot on the lock. She throws some fresh wood into the fire, then relights the fire. Olaf asks where Hans has been, and Anna tells him that it was not true love. They run all the way there, but they cannot find any other act of "true love" to save her. In the meantime, a cold wind blows in through one of the windowpanes--Kristoff and Sven are coming back. Soon, however, the walls of the castle crack under the pressure of the ice, forcing them to scramble out of the room. Over the course of the next few days, both Elsa and Anna struggle down the hallway, but are almost blinded from the snow and strong wind. As they try to get to the bottom of the hall, an ice spike grows around them, blocking their path. At first, they manage to slide down the building, but eventually, they reach the top as a huge snowball. Later, during a terrible storm, Sven and Kristoff fight through the icy ice. When a ship rises by ice and capsizes them, the mast smashes, and the entire ship sinks. It is too wide for Sven to cross, so he slips into the freezing water below. After a moment, he calls out for his friend to go on, but before he can do so, Anna faints. Her sister, who has returned from the mountains, is dead because of her. With that, the storm stops. Everyone rushes toward the wall edge to see what has happened. Anna is still able to walk across the frojords, but now she can see across the city. Anna hears someone drawing a sword from its cabbard, and she turns to see Hans behind her. He raises his head over his head, and he draws his sword against her. Anna throws himself in front of Nemo's sister, whom she thinks is dead. However, when she realizes how much she has sacrificed herself for him, she begins to cry. Finally, she embraces her sister, saying, "Anna, you sacrificed myself for me? Anna, weak, good naturedly: I love you. You sacrificed yourselves for me! Anna says, "I love you
|
244 |
+
|
245 |
+
---
|
246 |
+
|
247 |
+
Section Scores for script_frozendisney_summary:
|
248 |
+
|
249 |
+
- -1.3658
|
250 |
+
|
251 |
+
- -1.498
|
252 |
+
|
253 |
+
- -1.3065
|
254 |
+
|
255 |
+
---
|
256 |
+
|
257 |
+
## script_strangersonatrain_summary
|
258 |
+
|
259 |
+
The action shifts to a train station in Washington, DC. A young man wearing expensive tennis shoes is being taken out of a taxi and into a parlor car where he meets Guy Haines, a well-dressed tournament tennis player who has just returned from college. When Guy asks about Anne Burton's divorce, the young man responds that she was the daughter next door. He also mentions that he will be returning to Washington tomorrow for a doubles match. As they are finishing their lunch, Guy tells Bruno that his father hates him because he thinks he should get rid of his wife before he dies. At this point, Guy decides to go to "Metcalf," a town on the northeast coast of Virginia, to talk with Anne. They discuss the fact that each of them has a person he would like to kill, but can't kill the one he needs to get rid off. Each murderer does the other's murder, then there is no connection between them; the one who did the motive is not there. After the two men leave, Guy picks up his lighter and starts to call it to him.
|
260 |
+
Back to Scene One, the Senator tells Anne and Guy that they are free from a scandal. He reminds them that even the "most unworthy" of them can have the right to live and pursue happiness. The scene shifts to the police station in Washington, D. C. where Guy is waiting for Captain Turley at his office. He meets Professor Collins, the drunken companion who was with Guy on the train the previous night. Guy asks whether he remembers him, and Collins says that he doesn't remember anything about the trip from New York. Guy wonders aloud how he could have missed the train when he didn't know what he was supposed to be talking about. A man walks up and down the street across the street, looking for Guy. Guy recognizes this man as Leslie Hnnessy whom he has been working for sixteen hours every day. When Guy returns, he finds a note under the door, which reads: "Call me at Arlington." Guy looks at the note and realizes that it is handwritten by a Frenchman named Antony. They walk out of the painting, and Guy sees a small figure standing behind a tall white column near the spot where they had disappeared. Guy calls out to the figure, but Anne does not turn away. She notices that the figure is a tennis fan. Two shots later, Guy approaches the figure and introduces himself as Antony, an acquaintance of Mr. and Mrs. Darville. After Guy leaves, Hammond enters and takes over the role of Guy's assistant. At the end of the scene, the guests arrive at the Burgomaster's house. In the meantime, there is talk of a murder being carried out by one of the members of the party. It turns out that the two men are discussing the idea of using the life force to make atomic electricity look like a buggy. As the guests leave, Madame Cunningham joins them. She wants to use a poison to kill her husband, but Guy refuses to give her a gun. This leads to a violent confrontation between the two women. Everyone rushes off to get help, except for Barbara, who has fainted. On the way back to the house, Guy knocks on the door and pulls Bruno back onto the seat. Guy throws him into a deep trance, and then he pushes him back on the chair. Guy demands to know why he did it, and Anne begins to believe that she is responsible for the murder. Later, Guy decides to go to his father's home, so he sends Haines to pick up the key to his house.
|
261 |
+
The narrator tells us that Guy has been trying to get his lighter back from the island, which is why he dropped it in the middle of the night. He doesn't know what to do with it, so he decides to go up to Anne's room and change clothes. At the theater, Guy and Anne are sitting at a table discussing a game they're about to play. Guy asks Anne if she'd like to have her lighter put on the island before the police find it, but Anne refuses. She says she'll wait until after the third set of games to see if Guy can finish off the first set in time. Guy gets into a fight with Reynolds, who plays hard but holds his own. In the second set, Guy wins by a score of six-five. A passerby accidentally drops his cigarette lighter, which turns out to be Guy Haines's. As the spectators watch, Guy knocks the lighter off one of the ledges below, making it impossible for him to reach it. Finally, he manages to break through the crowd and win the game.
|
262 |
+
|
263 |
+
---
|
264 |
+
|
265 |
+
Section Scores for script_strangersonatrain_summary:
|
266 |
+
|
267 |
+
- -1.1219
|
268 |
+
|
269 |
+
- -1.3687
|
270 |
+
|
271 |
+
- -1.2305
|
272 |
+
|
273 |
+
---
|
274 |
+
|
275 |
+
## script_sunsetblvd._summary
|
276 |
+
|
277 |
+
"Sunset Boulevard" is a street in Los Angeles, California. A police squad car has been stationed at the scene of a murdered man's body floating in a swimming pool. In this scene, we are introduced to Joseph Gillis, an unmarried writer who lives in one of the most decrepit apartment buildings in the city. He works as a typewriter and sits on a dirty bed with a scattered pile of typewritten and pencil-marked pages. Two men come into the room and ask for the keys to the car they have loaned him. They do not believe them because they have a court order that they must return the car by noon tomorrow or there will be fireworks. As they leave, Gillis tells his story about a young shortstop who is trying to become a successful baseball player. Betty enters and takes a packet of papers from her desk. She puts it on Gillis' desk but doesn't notice Gillis. When she sees him, she gives him a copy of Plot 27A, which was written by James Joyce. After Gillis leaves, Skolsky greets him and offers to give him three hundred dollars so he can buy a new car. Morino approaches and asks Gillis for three hundred bucks. Gillis refuses, saying he needs more money than ten percent of his income. Morinos attempts to get Gillis to sit behind the typewriter, but Gillis says he cannot. The next time he goes back to Hollywood, he stops at Norma Demond's house, where he meets Max van Mayerling, a sixty-year-old man who is all black except for his shirt, collar, and bow tie. This makes Gillis wonder why he hasn't kept him waiting so long since he had landed himself in the flat tire; he also realizes that he couldn't walk back to his apartment until he got out of the driveway. On the way back to the house, however, Matt Mingott shows up and helps Gillis remove his car from the garage. While Gillis waits, Madame Desmond comes down the stairs and introduces him to Norma desmond, whom she calls "the lady's voice." She leads Gillis upstairs to a huge living room filled with faded photographs and a pipe organ. It is here that Gillis begins to write a script. He does not want to read it, but he wants to find someone who can help him. He decides to take the manuscript home and finish it while he does a patch of the script. Later, when Gillis is gone, Mrs. Desmond invites him over to spend the afternoon with her.
|
278 |
+
The narrator tells us that Gillis and Norma are going to have a party, but they're not going to share it with the other guests. Nora asks Gillis what time they should get there, and he says he doesn't know. He thinks someone bought up all of the tickets for "a performance" at the Metropolitan in order to sit there listening to Le Traviata while he was afraid of getting cold. No one else is there, so they don't need to share this party with anyone else. As they dance, Gillis realizes that he has fallen in love with her. He gives her a box filled with music and a case with a gold cigarette and lighter. When she reads the contents, he opens the box and engraves some notes on it: "To Joe from Nora," "two bars of music" and "Madam about the boy". She rushes into her room and bangs down the door. At the end of the chapter, we see Gillis taking a walk through the rainy night. He goes to Green's apartment, where he meets Artie Green who is conducting a choral with his ladle. Artie introduces himself to Gillis as an opium-smuggler named Joe Gillis. Gillis invites Artie to stay for a few weeks, and Artie agrees. Meanwhile, Betty enters the room and finds Gillis waiting for her. They talk about their old stories, and Gillis tells her about how he had a schoolteacher like that when he first came to live in Burmese. Now he wants to start writing a new story. But he can't bring himself to do anything until he promises to behave like a decent human being. In another part of the house, Northumberland's wife tries to get Gillis to deliver the script to de Mille, but he is nowhere to be found. It turns out that De Mille didn't approve of Gillis because he thought he could make a star out of him. This makes Gillis even more determined to finish the script.
|
279 |
+
Betty is on her way home from work when she receives a strange call from a woman named Betty. She tells her that Gillis lives in the same house as her, and she asks him to come over and see it for herself. When he does, she realizes what's going on and goes back to her room where she sits staring at Gillis. Then she hears the doorbell ring, so she climbs up onto the bed and finds a gun hidden under her pillow. As she raises herself to reveal it, she dismounts her pillow and hides the revolver beneath it. At the end of the scene, Joe comes into the house and leads Betty inside. He shows her around the house, telling her that there are eight master bedrooms, a sunken tub, an alley with a swimming pool, and a movie theater. He also gives her some clues about how Norma came to live in such a small house. Before they leave, Gillis leads her to the upper landing, where she lies down on the floor watching him. After he leaves, she crawls towards a console and removes the patches on her face and then goes to his room. There, she takes off the patches and puts them off her face. She enters Gillis's room and begins to take care of him. In the middle of her conversation with Gillis, she says that she doesn't think she has the courage to confront the people who want her. They don't want her; they just want her because she gets letters every day. No matter how much money she wants, she won't be allowed to do anything without him. Finally, she opens the door and rushes out. On the other side of the pool, policemen and detectives are trying to stop a mob. A lieutenant walks through the crowd and calls on the coroner. It turns out that the head of Holmby Hill's division and the lad's homicide squad have been talking to each other. They can't deny having shot Gillis--they didn't even know if he was a murderer or whether he wanted to blackmail her. Meanwhile, the newsreels have arrived with their cameras, and they set up the stairs to the palace.
|
280 |
+
|
281 |
+
---
|
282 |
+
|
283 |
+
Section Scores for script_sunsetblvd._summary:
|
284 |
+
|
285 |
+
- -1.3195
|
286 |
+
|
287 |
+
- -1.3021
|
288 |
+
|
289 |
+
- -1.2669
|
290 |
+
|
291 |
+
---
|
292 |
+
|