Skip to content
MyDSHMy DSH Field Notes
中文
Independently maintained; not affiliated with DeepSeek. No association with mydsh.dev or other MyDSH sites.

All categories/Translation and localisation

Translation practice · Level Advanced · Translation MT Post-editing

Post-editing MT: what to fix

Machine output as a base needing human editing.

Prompt

First decide what the text does: inform, persuade or process. The same sentence translates differently. You translate function, not words.

Task: Machine output as a base needing human editing.

[Before you start, do three things]
1. Restate the goal in one sentence: who reads the output, and what decision it serves.
2. If I have not given you the following, ask me. Do not assume:
   - text type and quality bar
   - MT engine
   - glossary
   - time budget
3. Say how you plan to do it (at most 5 steps) and wait for my go-ahead.

[Go through each of these]
- decide light or full post-editing
- fix errors by category
- prioritise negation, conditionals, references
- judge whether retranslating is faster

[Process]
1. Confirm purpose, audience, venue
2. Fix terminology and proper nouns
3. Restructure per target-language habit
4. Flag what cannot be translated literally and needs the author
5. Check numbers, units, names

[Output]
Translation, glossary, list of uncertainties.

[Do not]
- No word-for-word rendering
- Never smooth away facts or qualifiers
- No inconsistent invented terminology
- Never present guesses as fact

What to state before it starts

Leave these unsaid and the model fills them with an assumption. Most 'wrong answers' start here.

  • text type and quality bar
  • MT engine
  • glossary
  • time budget

Go through each of these

Where this kind of work most often goes wrong. Already in the prompt, so you do not have to remember it.

  • decide light or full post-editing
  • fix errors by category
  • prioritise negation, conditionals, references
  • judge whether retranslating is faster