ChatGPT Prompt Engineering: Master Productivity with The “Sequence Prompt”

TL;DR

Use a sequence prompt that generates a numbered table of 10 improvement options for the provided text, then asks the user to pick one.

Briefing Cornell Notes

Briefing

A “sequence prompt” workflow turns one-shot editing into a controlled, choice-based editing loop: it generates a numbered menu of rewrite options, asks which improvement to apply, and maintains a change log so users can track exactly what was altered. The practical payoff is speed with accountability—writers can iterate on a text without losing sight of why each revision happened or what changed.

The prompt is built in stages. First, it includes instructions (in brackets) telling the model to work on a piece of text provided inside curly brackets. Next, it generates a table with 10 distinct improvement suggestions, numbered in the left column so a user can select an option. After the table appears, the prompt asks a follow-up question—“what improvements would you like to make to the text pick one from the table above”—and then uses an “acknowledge with dot dot dot” step to delay table creation until the model has the full context it needs.

In the Matrix example, the user pastes a snippet and requests edits. The output returns a table listing 10 improvement ideas, such as clarifying the relationship between free will and fate, adding context about the Matrix trilogy, expanding the Oracle’s role, addressing narrative inconsistencies, and shifting to a more formal tone. Crucially, each option is paired with a “change log” entry describing what will be implemented. When the user selects improvement number three, the change log specifies adding background on the trilogy—director, release date, story summary, and key movie identifiers—plus adjusting tone and reducing direct quotes. Selecting another option later (like number 10) adds a conclusion summarizing the main points, demonstrating how the same structure supports multiple rounds of refinement.

The workflow also adapts beyond fiction. With a mental-health style input (“I don’t feel good mentally… I’m anxious all the time”), the table proposes actionable self-care ideas (exercise, journaling), social support (finding a support group or community), and language adjustments for clarity. The change log reflects higher-stakes guidance too, including encouraging professional help (talking to a therapist or counselor). The user then selects specific improvements—adding concrete context about the anxiety (e.g., an upcoming job interview), followed by mindfulness guidance focused on the present moment, and finally a reminder that small steps still count as progress.

A third demonstration uses a poem. The table offers craft-level edits: clarifying message and stance, adding imagery and figurative language, adjusting rhythm scheme for variety, reworking flow, shortening or lengthening, and adding a final stanza for closure. After choosing options, the change log records what was modified—such as shortening by removing repetitive lines, changing the rhythm scheme, inserting imagery (including a metaphor about a key), and appending a concluding stanza. The overall pattern is consistent: pick a numbered improvement, apply it, and review the logged changes—making iteration feel less like guesswork and more like versioned editing.

Cornell Notes

The sequence prompt creates a repeatable editing loop: it generates a numbered table of 10 rewrite options for a provided text, asks which option to apply, and records a “change log” describing what each selected improvement will do. This structure helps users iterate without losing track of revisions. In the Matrix example, selecting options adds trilogy background, shifts tone, reduces quotes, and later appends a conclusion. With anxiety advice, the same mechanism produces practical self-care suggestions, encourages professional support, and adds specificity (like a job interview trigger) plus mindfulness and small-step encouragement. For a poem, it supports craft edits such as imagery, rhythm changes, shortening/lengthening, and adding a final stanza, all with logged changes.

How does the sequence prompt manage step-by-step editing instead of one-pass rewriting?

It uses a staged instruction structure: the prompt includes instructions (in brackets) and the target text (in curly brackets), then generates a table of 10 numbered improvement suggestions. After the table, it asks which improvement to apply. The “dot dot dot” acknowledgement step delays final table creation until the model has the needed context, and the workflow repeats by selecting another numbered option in later rounds.

What does the “change log” accomplish during revisions?

The change log records the specific modification tied to the selected option. For example, when improvement number three is chosen for the Matrix snippet, the log describes adding background on the Matrix trilogy (director, release date, story summary, and key movie identifiers), shifting toward a more formal/academic tone, and reducing direct quotes. When improvement number 10 is selected, the log reflects adding a conclusion summarizing the main points.

How does the workflow handle different input types, like advice vs. fiction vs. poetry?

The same table-and-selection mechanism works across domains. For anxiety advice, options include self-care (exercise, journaling), social support (support group/community), clarity improvements, mindfulness focused on the present moment, and encouragement to seek professional help (therapist/counselor). For poetry, options target craft elements like imagery/figurative language, rhythm scheme changes, shortening/lengthening, and adding a final stanza for closure.

Why is selecting a specific numbered improvement useful compared with asking for a rewrite directly?

Numbered options turn editing into controlled choices. Instead of requesting “rewrite this,” the user can pick a targeted improvement—like expanding the Oracle’s role, adding trilogy context, or changing the poem’s rhythm scheme—and then see exactly what will be implemented via the change log. This makes iteration more transparent and easier to manage.

What kinds of improvements appear in the poem example, and how are they reflected in the change log?

The poem table includes craft edits such as clarifying the message and stance, adding specific examples, addressing concerns, adding imagery and figurative language (e.g., a metaphor about a key), adjusting rhythm scheme (the transcript shows an AABBCCDD-style pattern), shortening by removing repetitive lines, and adding a final stanza for closure. Each chosen option is logged with a description of what changed.

Review Questions

If a user wants to revise a text multiple times, how does the prompt’s table + selection + change-log structure support that workflow?
In the Matrix example, what specific categories of edits appear in the numbered suggestions, and how does the change log describe them?
For the poem task, which improvement types target structure (rhythm/flow/stanzas) versus meaning (message/imagery), and how would you choose between them?

Key Points

1
Use a sequence prompt that generates a numbered table of 10 improvement options for the provided text, then asks the user to pick one.
2
Include a “change log” so each selected option records exactly what will be implemented in the revised text.
3
Iterate by selecting different numbered improvements across multiple rounds rather than requesting a full rewrite each time.
4
The same workflow can handle fiction snippets, mental-health style advice, and poetry by producing domain-appropriate suggestions.
5
For fiction, options can include tone shifts, added context (e.g., trilogy background), narrative consistency fixes, and structural additions like conclusions.
6
For advice, options can include practical self-care, mindfulness prompts, specificity about triggers, and guidance to seek professional support when appropriate.
7
For poetry, options can target imagery, figurative language, rhythm scheme, length, flow, and adding a concluding stanza.

Highlights

The prompt turns editing into a choice-based loop: generate 10 options, select one, apply it, and log the change.

Selecting improvement number three in the Matrix example adds trilogy background details (director, release date, story summary, key movie identifiers) and shifts tone.

In the anxiety example, the change log includes both practical coping ideas and encouragement to seek professional help (therapist/counselor).

For the poem, the workflow can adjust craft elements like rhythm scheme and add a final stanza, with each edit recorded in the change log.

Topics

Sequence Prompt
Prompt Engineering
Change Log
Iterative Editing
Poem Rewriting