#英语综合测（一）

• 定位：学完词汇、长难句、阅读、完形、翻译、写作六个板块一轮之后的科目级混合检验——检验跨板块调用能力：完形里考词汇与逻辑衔接、阅读里考长难句定位、翻译里考拆解输出、写作考整体组织，一卷打通六个板块
• 结构：A 完形 1 篇 15 空（每空 0.5 分，共 7.5 分）+ B 阅读理解 2 篇各 4 题（每题 2 分，共 16 分）+ C 翻译 2 段（每段 5 分，共 10 分）+ D 写作 1 题（20 分），满分 53.5 分；建议计时 105 分钟（完形 15 / 阅读 32 / 翻译 18 / 写作 40）
• 板块对应：词汇见 词汇策略与分层记忆、句法基础见 长难句拆解、阅读方法见 阅读理解题型与解题框架、写作见 写作框架与素材库、完形见 完形填空、翻译见 翻译
• 要求：全部文章与题目为模拟题（真题风格自编，不与任何真题原文重合）。答案与解析在独立答案页 04-英语综合测-答案，做完再对
• 考后动作：错题登记进 00-错题本——完形错因标签"词义/逻辑/搭配"，阅读错因标签"定位/推理/词义/主旨"，翻译按采分点自评失分句型，写作对照框架自评段落功能缺失项

#A 部分：完形填空（15 空）

Directions: Read the following text. Choose the best word for each numbered blank.

Habits are the brain's way of 1 effort. When a behavior is repeated in a stable context, the mental work it demands drops sharply: what once required 2 attention becomes nearly automatic. Psychologists describe this as a shift from goal-directed control, in which outcomes are constantly 3, to habitual control, in which cues in the environment 4 the behavior directly.

The transition has a price. Once a habit is 5 in place, it survives even when its original purpose disappears. A person who once checked a phone for urgent messages may keep 6 it hours after anything urgent has stopped arriving. The behavior is no longer 7 by its consequences; it is triggered by the situation itself.

That is 8 context matters so much in habit change. Interrupt the cue and the habit loses its 9; keep the cue and willpower alone rarely 10. Studies of people who move house, 11, find that their habits—good and bad alike—are unusually changeable, because the old cues have been 12 away.

The practical lesson is not to fight habits with determination but to 13 them: attach a desired behavior to an existing cue, keep the 14 small at first, and let repetition do the rest. Behavior change, in this view, is less a 15 of character than a matter of design.

第 1-5 题

1. A. wasting　B. saving　C. paying　D. spending
2. A. delicate　B. desperate　C. deliberate　D. detached
3. A. neglected　B. suspended　C. exaggerated　D. monitored
4. A. forbid　B. justify　C. trigger　D. postpone
5. A. loosely　B. firmly　C. rarely　D. randomly

第 6-10 题

6. A. to check　B. checked　C. check　D. checking
7. A. rewarded　B. convinced　C. driven　D. measured
8. A. when　B. why　C. how　D. whether
9. A. grip　B. shape　C. pace　D. value
10. A. fails　B. works　C. counts　D. differs

第 11-15 题

11. A. in addition　B. by contrast　C. for instance　D. as a result
12. A. taken　B. put　C. broken　D. swept
13. A. resist　B. ignore　C. redesign　D. recall
14. A. changes　B. rewards　C. goals　D. risks
15. A. virtue　B. matter　C. product　D. crisis

#B 部分：阅读理解（2 篇，各 4 题）

#Text 1（主题：人工智能的可解释性）

In 2018, researchers at a major hospital showed that a deep-learning system could diagnose certain eye diseases from retinal scans as accurately as senior specialists. The system was never asked to explain itself, and it never did. Its judgments emerged from millions of tuned parameters, none of which carries meaning on its own. For the doctors who used it, this silence was not a technical detail but a practical problem: a diagnosis they cannot interrogate is a diagnosis they cannot fully trust.

The field has since produced "explainability" tools that highlight the parts of an image that most influenced a model's decision. These heat maps are popular and easy to read, but critics warn that they may explain nothing at all. A map can show where a model looked without showing what it saw; more importantly, different tools applied to the same model often produce conflicting maps, which suggests that at least one of them is describing the model inaccurately.

The deeper difficulty is conceptual. Human explanations are selective: a doctor justifies a diagnosis by citing a few decisive features, not by replaying the full neural computation that produced it. We accept this selectivity because we share assumptions about what counts as a good reason. Machines do not participate in that shared culture, so an "explanation" generated to satisfy us may be less a report of the model's actual reasoning than a plausible story shaped for human consumption.

None of this argues against using such systems. In many settings, a well-tested model deserves trust even if it cannot explain itself—airline autopilots have earned decades of confidence without offering passengers any reasons. But the appropriate response to an unexplainable system is validation: extensive trials, continuous monitoring, and honest disclosure of failure rates. Trust, in other words, can rest on evidence about performance rather than on access to reasoning. What it cannot rest on is the comforting but unearned feeling that a colorful map has let us see inside the machine.

16. According to Paragraph 1, the doctors' problem with the eye-disease system was that ______

A. its accuracy fell short of that of senior specialists B. it could not answer questions about its own judgments C. it demanded far more data per scan than expected D. its parameters were too few to be properly tuned

17. Critics' main concern about the heat-map tools (Paragraph 2) is that ______

A. they are too difficult for doctors to read B. they are rarely applied to the same model twice C. they may misrepresent the model's actual behavior D. they highlight image regions that did not influence the decision

18. The author suggests that a machine-generated "explanation" may become "a plausible story" (Paragraph 3) because ______

A. machines do not share humans' assumptions about what counts as a good reason B. machines deliberately hide their reasoning from human users C. human doctors refuse to accept reasons offered by machines D. neural computation cannot be described in mathematical language

19. Which of the following best states the main point of the text?

A. Explainability tools have made medical AI fully trustworthy. B. Unexplainable systems should be withdrawn from clinical use. C. Trust in opaque systems should rest on tested performance rather than the illusion of explanation. D. Heat maps will eventually replace clinical trials as evidence of safety.

#Text 2（主题：四天工作制试验）

The results of the world's largest trial of a four-day working week were published recently: across sixty-one companies, most kept the shorter schedule after the trial ended, and average revenue held steady. Enthusiasts declared the debate settled. It is not.

The trial's design limits what it can tell us. Companies joined voluntarily, which means the sample was weighted toward employers already sympathetic to the idea—firms confident that their work could be compressed. Sectors where output is tied directly to hours, such as hospitals and factories, are largely absent. And a six-month window cannot reveal effects that mature slowly: careers built on visibility, the slow accumulation of mentoring, the maintenance of client relationships that depend on presence.

There is also a distributional question the headlines ignored. When office workers compress five days into four, someone still must cover the shifts that no longer fit. The burden tends to land on those with the least bargaining power—part-timers, junior staff, contractors—while the flexibility dividend flows to those already well positioned. A reform advertised as universal may quietly widen gaps.

None of this makes the four-day week a bad idea. The honest conclusion is narrower: in some kinds of work, for some workers, a compressed schedule can hold output constant while improving well-being, at least over the short term. Whether that finding travels beyond motivated volunteers and knowledge-work offices is exactly what better experiments must now test. Treating a promising but limited trial as a verdict does not accelerate reform; it only makes the eventual disappointment louder.

20. One limitation of the trial mentioned in Paragraph 2 is that ______

A. participating companies had chosen to join the trial B. the trial lasted only sixty-one weeks in total C. average revenue declined during the trial period D. hospitals were forced to cut their working hours

21. The "effects that mature slowly" (Paragraph 2) refer to things such as ______

A. quarterly fluctuations in company revenue B. careers and client relationships built on continued presence C. immediate gains in employee satisfaction D. extra shifts covered by part-time workers

22. The "distributional question" (Paragraph 3) implies that the compressed schedule ______

A. benefits all workers equally across sectors B. shifts its costs onto workers with the least bargaining power C. mainly helps part-timers, junior staff and contractors D. removes the need for shift coverage altogether

23. The author's attitude toward the four-day week trial's findings is one of ______

A. firm opposition　　B. unqualified enthusiasm C. cautious reservation pending broader evidence　　D. complete indifference

#C 部分：翻译（2 段，英译汉）

Directions: Translate the following passages into Chinese. Write your translation on the ANSWER SHEET.

第 46 段

The technologies that will shape the next decade are being built now, often by people who assume that their effects can be sorted out later. History suggests otherwise. Every major technology, from the printing press to the automobile, reshaped society in ways its inventors neither intended nor foresaw; what distinguished successful adaptations was not cleverer machines but faster social learning.

第 47 段

Sleep is not a passive shutdown but an active maintenance cycle. During deep sleep, the brain clears metabolic waste that accumulates during waking hours; during rapid-eye-movement sleep, it appears to replay and reorganize the day's experiences, strengthening some memories and discarding others. A chronic shortage of sleep, therefore, is less like an inconvenience than like skipping maintenance on a machine that runs twenty-four hours a day.

#D 部分：写作（1 题）

第 48 题（图画作文，160-200 词）

Write an essay of 160-200 words based on the following picture. In your essay, you should

1. describe the picture briefly,
2. interpret its intended meaning, and
3. give your comments.

图画内容（文字描述）：一位年轻人站在两条岔路口前。左边的路标写着"捷径"，路上看似平坦但通向迷雾；右边的路标写着"积累"，路口立着一块小牌子"日拱一卒"，道路蜿蜒但远处可见山顶。年轻人背包上写着"备考"。

You should write neatly on the ANSWER SHEET.（20 分）

#计分与测后统计（自填）

错题逐条回写 00-错题本；完形错题词与阅读定位错句进入次日顽固清单。跨板块诊断：完形错在逻辑衔接（3/7/8/11/13）而阅读主旨题也错，说明篇章结构意识不足，回读 阅读理解题型与解题框架 的"段落功能分析"节；翻译里 less...than/not...but 类结构与阅读 Text 1 若同时失分，回读 长难句拆解 的比较结构专题。

答案与解析见独立答案页：04-英语综合测-答案（独立成页便于打印试卷时不含答案）。