Rule reference · Efficiency

Filler phrases

  • filler_phrases
  • info in this example
  • Efficiency
  • stage · Redundancy detection

Fires when: Politeness and intensifiers that do not change the task.

Politeness does no harm to quality, but in a prompt that runs thousands of times a day it is a measurable cost with no effect on the output.

The engine removes filler automatically in balanced mode and verifies that every requirement survives.

Example that triggers it

Hi! I would like you to please summarize the report below in 3 bullet points. Thank you so much!

<report>
{{report}}
</report>

info 4 filler phrases that don't change the instruction.

Why it matters. Politeness and intensifiers cost tokens on every call without changing what the model is asked to do.

Quality 97/100 · 32 tokens · 1 findings

Fixed version

Summarize the report below in 3 bullet points.

<report>
{{report}}
</report>

Quality 100/100 (+3) · 21 tokens · 0 findings

Resolved by the fix

  • filler_phrases: 4 filler phrases that don't change the instruction.

Scores, token counts (GPT-4.1 tokenizer), and findings on this page are computed by the engine when the site is built.

How it affects the score

Each finding subtracts a fixed penalty from 100: critical 30, warning 10, info 3. This rule counts against the Efficiency dimension. See the scoring model for the full formula.

Checked in the same pipeline stage

All 20 rules →