# Why context matters in chat moderation

> The same sentence can be harmless or harmful depending on what came before. How conversation context changes moderation, with real examples.

Source: https://www.vigimod.com/en/learn/context-in-moderation · VigiMod (https://www.vigimod.com) · updated 2026-11-11 · 5 min

## In short

- Many messages only make sense after the ones before them: replies, jokes, spoilers.
- Without context a model guesses; with context it reads the situation.
- Context should help to read the message, not move the blame onto it.
- Sending the last 5–10 messages is usually enough.

## One sentence, two meanings

| Previous messages | Message | Without context | With context |
|---|---|---|---|
| “anyone seen the finale?” – “not yet, tonight!” | “lol he dies at the end” | harmless | **spoiler** |
| “my goldfish is old, I'm worried” | “I think he dies at the end, sorry” | spoiler? | **harmless** |
| “you lost again 😂” – “rematch?” | “you're going down” | threat? | **friendly banter** |
| “stop writing to me” – “I said stop” | “I know where you live” | ambiguous | **threat** |

The message alone does not decide. The conversation does.

## What context fixes

- **False alarms** on jokes between friends, gaming banter, replies that quote someone.
- **Rules that depend on the topic:** spoilers, off-topic messages, selling in the wrong channel.
- **Escalation:** a message that is mild alone but is the third unwanted contact in a row.

## What context must not do

The score is about **the last message**. If someone else insulted you two messages ago, your polite reply must not be flagged as an insult. A good model uses context to *understand* the message, not to transfer the previous messages' harm onto it.

## How much context?

In chat, meaning usually depends on the last few messages. Sending 5–10 previous messages, oldest first, covers almost every case while keeping requests fast. Very long histories cost more and rarely change the verdict.

## Why it is hard for models

Context-aware training data is rare: most public datasets contain single messages. Models must see many examples where the conversation **changes** the right answer, in both directions, to learn to use it. That is one of the main goals of the next VigiMod version.

## In VigiMod

Send the previous messages in `context`, oldest first, up to 10. The scores remain about `text`; the context only helps to read it. See the [documentation](https://www.vigimod.com/en/docs) and the [Discord guide](https://www.vigimod.com/en/discord) for an example that fetches the last messages of the channel.

## Quiz

1. “lol he dies at the end” after “anyone seen the finale?” is…
   Answer: a spoiler. The conversation is about a series, so it reveals the plot.
2. With context, the score should be about…
   Answer: the last message only, read with the help of the previous ones. Context helps to read the message, it does not transfer blame.
3. How many previous messages are usually enough in a chat?
   Answer: 5–10. Meaning usually depends on the last few messages.
4. Why is context hard for models to learn?
   Answer: Few datasets contain conversations where context changes the answer. Most public data has single messages only.