Corpus Viewer

Work: clean Viewing: Source

Create annotation system (creates a ticket)

Manual annotation is the default. Approved files are written as sibling pages under kanbun/, hanmun/, or hanvan/ with the same filename as this source page.

Annotation system

Summary

Reasoning

Digital keyboard

Kanbun-LM is not shown here because automatic generation is Japanese-only. This patch keeps the manual path as the default workflow.

Source page


# WORK_TITLE: 大語言模型
# DISPLAY_TITLE: 大語言模型
# AUTHOR:
# TIMES:
# PAGE_TITLE: 大語言模型
# CATEGORIES:
大語言模型，語言模型之長也，基於深度學習之術。集海量文辭，納古今中外之言，融彙百家，煉成一身。可作文賦詩，解疑答惑，譯諸國之言
。

蓋模型之所以大，因其參數衆多，動輒數十億乃至千億，內存浩瀚如海，包羅萬象。其訓練源材，涵蓋經
史子
集、
詩詞
歌賦、
小説戲曲、
論文報告、網絡文摘，乃至居處之對話，無所不有。經年纍月，吸納衆長，漸成一家之言
。

然模型亦有不足之處，其所習皆爲舊聞，難以時新，故於時事變遷或有未逮。且模型之言，蓋爲模式匹配，淺不足與測深。

據

Annotation-system text

Material note *

Provenance *

Choose every label that applies. At least one is required.

User-made Public domain Historical source or manuscript

References

A formal citation style is welcome but not required.

AI assistance disclosure

AI-assisted What was assisted?

Evidence links

Links are preferred when the evidence already exists online.

Evidence files

Up to 10 files and 20 MB for this ticket. HTML, XHTML, and SVG are rejected.

Optional contact details

Stored separately and encrypted. Deleted after 30 days or when the ticket closes.

Name Email Contact note

Save this ticket key on this device

Unchecked by default. The key is otherwise shown once and is not stored by the browser.

Ticket ID:

Ticket key (save this): Open ticket

Add companion material (creates a ticket)

Approved material appears beside this corpus text. The 20 MB limit applies to this submission only; the text may accumulate any number of approved submissions.

Material type

Optional edition or material label

Translations are identified in menus by language and translator, not by this label.

Translation language (ISO 639-3)

The stored value is the selected three-letter ISO 639-3 code.

Translator

Leave blank for anonymous or unidentified translations.

Translation text

Material files

PNG, JPEG, or PDF. Together with evidence files, this ticket may contain up to 10 files and 20 MB.

Related corpus path

Material links

Ticket summary

Reviewer reasoning

Material note *

Provenance *

Choose every label that applies. At least one is required.

User-made Public domain Historical source or manuscript Author-provided: the work's author supplied this translation

References

A formal citation style is welcome but not required.

AI assistance disclosure

AI-assisted What was assisted?

Evidence links

Links are preferred when the evidence already exists online.

Evidence files

Up to 10 files and 20 MB for this ticket. HTML, XHTML, and SVG are rejected.

Optional contact details

Stored separately and encrypted. Deleted after 30 days or when the ticket closes.

Name Email Contact note

Save this ticket key on this device

Unchecked by default. The key is otherwise shown once and is not stored by the browser.

Ticket ID:

Ticket key (save this):

Propose a text edit (creates a ticket)

This does not change the corpus directly. It creates a ticket with a unified diff attached.

Title

Summary

Reasoning

Edited text

Evidence links

Links are preferred when the evidence already exists online.

Evidence files

Up to 10 files and 20 MB for this ticket. HTML, XHTML, and SVG are rejected.

Optional contact details

Stored separately and encrypted. Deleted after 30 days or when the ticket closes.

Name Email Contact note

Save this ticket key on this device

Unchecked by default. The key is otherwise shown once and is not stored by the browser.

Ticket ID:

Ticket key (save this):

Punctuation options

Preset Source Modern CJK (Traditional) Modern CJK (Simplified / PRC) Pure classical Strip Custom

Quotes

Family

Chinese quotation marks 「」『』 Latin speech marks “ ” ‘ ’ Chinese speech marks ＂＇ Off

Ordering

Traditional: 「甲『乙』丙」 Simplified: 『甲「乙」丙』 Vertical quote forms (﹁﹂﹃﹄)

Question / Exclamation Keep ？ Collapse ? → 。

Keep ！ Collapse ! → 。

Semicolon / Colon Keep ； Collapse ; → 。

Keep ： Collapse : → 。

Comma Use ， Use 、

Repetition mark No repetition mark (default) 々 (Horizontal) 〻 (Vertical) 二 (Western Zhou-style)

# WORK_TITLE: 大語言模型 # DISPLAY_TITLE: 大語言模型 # AUTHOR: # TIMES: # PAGE_TITLE: 大語言模型 # CATEGORIES: 大語言模型，語言模型之長也，基於深度學習之術。集海量文辭，納古今中外之言，融彙百家，煉成一身。可作文賦詩，解疑答惑，譯諸國之言。蓋模型之所以大，因其參數衆多，動輒數十億乃至千億，內存浩瀚如海，包羅萬象。其訓練源材，涵蓋經史子集、詩詞歌賦、小説戲曲、論文報告、網絡文摘，乃至居處之對話，無所不有。經年纍月，吸納衆長，漸成一家之言。然模型亦有不足之處，其所習皆爲舊聞，難以時新，故於時事變遷或有未逮。且模型之言，蓋爲模式匹配，淺不足與測深。據