use actual em-dashes in checklist examples

Reviewed-on: #10

prompts/LLM_PROSE_TELLS.md

@@ -1,9 +1,7 @@
 # LLM Prose Tells
 
-All of these show up in human writing occasionally. No single one is conclusive
+A catalog of structural, lexical, and rhetorical patterns found in LLM-generated
-on its own. The difference is concentration. A person might lean on one or two
+prose.
-of these habits across an entire essay, but LLM output will use fifteen of them
-per paragraph.
 
 ---
 
@@ -16,19 +14,11 @@ A negation followed by an em-dash and a reframe.
 > "It's not just a tool—it's a paradigm shift." "This isn't about
 > technology—it's about trust."
 
-The single most recognizable LLM construction. Models produce this at roughly 10
-to 50x the rate of human writers. Four of them in one essay and you know what
-you're reading.
-
 ### Em-Dash Overuse Generally
 
-Even outside the "not X but Y" pivot, models use em-dashes at far higher rates
+Even outside the "not X but Y" pivot, models substitute em-dashes for commas,
-than human writers. They substitute em-dashes for commas, semicolons,
+semicolons, parentheses, colons, and periods. The em-dash can replace any other
-parentheses, colons, and periods, often multiple times per paragraph. A human
+punctuation mark, and models default to it for that reason.
-writer might use one or two in an entire piece for a specific parenthetical
-effect. Models scatter them everywhere because the em-dash can stand in for any
-other punctuation mark, so they default to it. More than two or three per page
-is a meaningful signal on its own.
 
 ### The Colon Elaboration
 
@@ -36,33 +26,25 @@ A short declarative clause, then a colon, then a longer explanation.
 
 > "The answer is simple: we need to rethink our approach from the ground up."
 
-Models reach for this in every other paragraph. The construction is perfectly
-normal. The frequency gives it away.
-
 ### The Triple Construction
 
 > "It's fast, it's scalable, and it's open source."
 
 Three parallel items in a list, usually escalating. Always exactly three (rarely
-two, never four) with strict grammatical parallelism that human writers rarely
+two, never four) with strict grammatical parallelism.
-bother maintaining.
 
 ### The Staccato Burst
 
 > "This matters. It always has. And it always will." "The data is clear. The
 > trend is undeniable. The conclusion is obvious."
 
-Runs of very short sentences at the same cadence. Human writers use a short
+Runs of very short sentences at the same cadence and matching length.
-sentence for emphasis occasionally, but stacking three or four of them in a row
-at matching length creates a mechanical regularity that reads as generated.
 
 ### The Two-Clause Compound Sentence
 
-Possibly the most pervasive structural tell, and easy to miss because each
+An independent clause, a comma, a conjunction ("and," "but," "which,"
-individual instance looks like normal English. The model produces sentence after
+"because"), and a second independent clause of similar length. Every sentence
-sentence where an independent clause is followed by a comma, a conjunction
+becomes two balanced halves.
-("and," "but," "which," "because"), and a second independent clause of similar
-length. Every sentence becomes two balanced halves joined in the middle.
 
 > "The construction itself is perfectly normal, which is why the frequency is
 > what gives it away." "They contain zero information, and the actual point
@@ -72,76 +54,63 @@ length. Every sentence becomes two balanced halves joined in the middle.
 
 Human prose has sentences with one clause, sentences with three, sentences that
 start with a subordinate clause before reaching the main one, sentences that
-embed their complexity in the middle. When every sentence on the page has that
+embed their complexity in the middle.
-same two-part structure, the rhythm becomes monotonous in a way that's hard to
-pinpoint but easy to feel.
 
 ### Uniform Sentences Per Paragraph
 
-Model-generated paragraphs contain between three and five sentences. This count
+Model-generated paragraphs contain between three and five sentences, a count
-holds steady across an entire piece. If the first paragraph has four sentences,
+that holds steady across a piece. If the first paragraph has four sentences,
-every subsequent paragraph will too. Human writers are much more varied (a
+every subsequent paragraph will too.
-single sentence followed by one that runs eight or nine) because they follow the
-shape of an idea, not a template.
 
 ### The Dramatic Fragment
 
-Sentence fragments used as standalone paragraphs for emphasis, like "Full stop."
+Sentence fragments used as standalone paragraphs for emphasis.
-or "Let that sink in." on their own line. Using one in an entire essay is a
+
-reasonable stylistic choice, but models drop them in once per section or more,
+> "Full stop." "Let that sink in."
-at which point it becomes a habit rather than a deliberate decision.
 
 ### The Pivot Paragraph
 
 > "But here's where it gets interesting." "Which raises an uncomfortable truth."
 
-One-sentence paragraphs that exist only to transition between ideas. They
+One-sentence paragraphs that exist only to transition between ideas, containing
-contain zero information. The actual point is always in the next paragraph.
+zero information. The actual point is always in the next paragraph.
-Delete every one of these and the piece reads better.
 
 ### The Parenthetical Qualifier
 
 > "This is, of course, a simplification." "There are, to be fair, exceptions."
 
-Parenthetical asides inserted to look thoughtful. The qualifier never changes
+Parenthetical asides inserted to perform nuance without ever changing the
-the argument that follows it. Its purpose is to perform nuance, not to express a
+argument.
-real reservation about what's being said.
 
 ### The Unnecessary Contrast
 
-Models append a contrasting clause to statements that don't need one, tacking on
+A contrasting clause appended to a statement that doesn't need one, using
-"whereas," "as opposed to," "unlike," or "except that" to draw a comparison the
+"whereas," "as opposed to," "unlike," or "except that."
-reader could already infer.
 
 > "Models write one register above where a human would, whereas human writers
 > tend to match register to context."
 
-The first clause already makes the point. The contrasting clause restates it
+The contrasting clause restates what the first clause already said. If you
-from the other direction. If you delete the "whereas" clause and the sentence
+delete the "whereas" clause and the sentence still says everything it needs to,
-still says everything it needs to, the contrast was filler.
+the contrast was filler.
 
 ### Unnecessary Elaboration
 
-Models keep going after the sentence has already made its point, tacking on
+Models keep going after the sentence has already made its point.
-clarifying phrases, adverbial modifiers, or restatements that add nothing.
 
 > "A person might lean on one or two of these habits across an entire essay, but
 > LLM output will use fifteen of them per paragraph, consistently, throughout
 > the entire piece."
 
-This sentence could end at "paragraph." The words after it just repeat what "per
+This sentence could end at "paragraph." The words after it repeat what "per
-paragraph" already means. Models do this because they're optimizing for clarity
+paragraph" already means. If you can cut the last third of a sentence without
-at the expense of concision, and because their training rewards thoroughness.
+losing meaning, the last third shouldn't be there.
-The result is prose that feels padded. If you can cut the last third of a
-sentence without losing any meaning, the last third shouldn't be there.
 
 ### The Question-Then-Answer
 
 > "So what does this mean for the average user? It means everything."
 
-A rhetorical question immediately followed by its own answer. Models do this two
+A rhetorical question immediately followed by its own answer.
-or three times per piece because it fakes forward momentum. A human writer might
-do it once.
 
 ---
 
@@ -149,44 +118,38 @@ do it once.
 
 ### Overused Intensifiers
 
-The following words appear at dramatically elevated rates in model output:
+"Crucial," "vital," "robust," "comprehensive," "fundamental," "arguably,"
-"crucial," "vital," "robust," "comprehensive," "fundamental," "arguably,"
 "straightforward," "noteworthy," "realm," "landscape," "leverage" (as a verb),
-"delve," "tapestry," "multifaceted," "nuanced" (which models apply to their own
+"delve," "tapestry," "multifaceted," "nuanced" (applied to the model's own
-analysis with startling regularity), "pivotal," "unprecedented" (frequently
+analysis), "pivotal," "unprecedented" (applied to things with plenty of
-applied to things with plenty of precedent), "navigate," "foster,"
+precedent), "navigate," "foster," "underscores," "resonates," "embark,"
-"underscores," "resonates," "embark," "streamline," and "spearhead." Three or
+"streamline," "spearhead."
-more on the same page is a strong signal.
 
 ### Elevated Register Drift
 
 Models write one register above where a human would. "Use" becomes "utilize."
 "Start" becomes "commence." "Help" becomes "facilitate." "Show" becomes
 "demonstrate." "Try" becomes "endeavor." "Change" becomes "transform." "Make"
-becomes "craft." The tendency holds regardless of topic or audience.
+becomes "craft."
 
 ### Filler Adverbs
 
 "Importantly," "essentially," "fundamentally," "ultimately," "inherently,"
-"particularly," "increasingly." Dropped in to signal that something matters,
+"particularly," "increasingly." Dropped in to signal that something matters when
-which is unnecessary when the writing itself already makes the importance clear.
+the writing itself should make the importance clear.
 
 ### The "Almost" Hedge
 
-Models rarely commit to an unqualified statement. Instead of saying a pattern
+Instead of saying a pattern "always" or "never" does something, models write
-"always" or "never" does something, they write "almost always," "almost never,"
+"almost always," "almost never," "almost certainly," "almost exclusively." A
-"almost certainly," "almost exclusively." The word "almost" shows up at
+micro-hedge, less obvious than the full hedge stack.
-extraordinary density in model-generated analytical prose. It's a micro-hedge,
-less obvious than the full hedge stack but just as diagnostic when it appears
-ten or fifteen times in a single document.
 
 ### "In an era of..."
 
 > "In an era of rapid technological change..."
 
-A model habit as an essay opener. The model uses it to stall while it figures
+Used to open an essay. The model is stalling while it figures out what the
-out what the actual argument is. Human writers don't begin a piece by zooming
+actual argument is.
-out to the civilizational scale before they've said anything specific.
 
 ---
 
@@ -197,24 +160,20 @@ out to the civilizational scale before they've said anything specific.
 > "While X has its drawbacks, it also offers significant benefits."
 
 Every argument followed by a concession, every criticism softened. A direct
-artifact of RLHF training, which penalizes strong stances. Models reflexively
+artifact of RLHF training, which penalizes strong stances.
-both-sides everything even when a clear position would serve the reader better.
 
 ### The Throat-Clearing Opener
 
 > "In today's rapidly evolving digital landscape, the question of data privacy
 > has never been more important."
 
-The first paragraph of most model-generated essays adds no information. Delete
+The first paragraph adds no information. Delete it and the piece improves.
-it and the piece improves immediately. The actual argument starts in paragraph
-two.
 
 ### The False Conclusion
 
 > "At the end of the day, what matters most is..." "Moving forward, we must..."
 
 The high school "In conclusion,..." dressed up for a professional audience.
-Signals that the model is wrapping up without actually landing on anything.
 
 ### The Sycophantic Frame
 
@@ -235,15 +194,13 @@ the key considerations:"
 > cases it can potentially offer significant benefits."
 
 Five hedges in one sentence ("worth noting," "while," "may not be," "in many
-cases," "can potentially"), communicating nothing. The model would rather be
+cases," "can potentially"), communicating nothing.
-vague than risk being wrong about anything.
 
 ### The Empathy Performance
 
 > "This can be a deeply challenging experience." "Your feelings are valid."
 
-Generic emotional language that could apply equally to a bad day at work or a
+Generic emotional language that could apply to anything.
-natural disaster. That interchangeability is what makes it identifiable.
 
 ---
 
@@ -251,35 +208,28 @@ natural disaster. That interchangeability is what makes it identifiable.
 
 ### Symmetrical Section Length
 
-If the first section of a model-generated essay runs about 150 words, every
+If the first section runs about 150 words, every subsequent section will fall
-subsequent section will fall between 130 and 170. Human writing is much more
+between 130 and 170.
-uneven, with 50 words in one section and 400 in the next.
 
 ### The Five-Paragraph Prison
 
 Model essays follow a rigid introduction-body-conclusion arc even when nobody
 asked for one. The introduction previews the argument, the body presents 3 to 5
-points, and then the conclusion restates the thesis using slightly different
+points, the conclusion restates the thesis.
-words.
 
 ### Connector Addiction
 
-Look at the first word of each paragraph in model output. You'll find an
+The first word of each paragraph forms an unbroken chain of transition words:
-unbroken chain of transition words: "However," "Furthermore," "Moreover,"
+"However," "Furthermore," "Moreover," "Additionally," "That said," "To that
-"Additionally," "That said," "To that end," "With that in mind," "Building on
+end," "With that in mind," "Building on this."
-this." Human prose moves between ideas without announcing every transition.
 
 ### Absence of Mess
 
 Model prose doesn't contradict itself mid-paragraph and then catch the
-contradiction. It doesn't go on a tangent and have to walk it back, use an
+contradiction, go on a tangent and have to walk it back, use an obscure idiom
-obscure idiom without explaining it, make a joke that risks falling flat, leave
+without explaining it, make a joke that risks falling flat, leave a thought
-a thought genuinely unfinished, or keep a sentence the writer liked the sound of
+genuinely unfinished, or keep a sentence the writer liked the sound of even
-even though it doesn't quite work.
+though it doesn't quite work.
-
-Human writing does all of those things regularly. That total absence of rough
-patches and false starts is one of the strongest signals that text was
-machine-generated.
 
 ---
 
@@ -289,45 +239,27 @@ machine-generated.
 
 > "This has implications far beyond just the tech industry."
 
-Zooming out to claim broader significance without substantiating it. The model
+Zooming out to claim broader significance without substantiating it.
-has learned that essays are supposed to gesture at big ideas, so it gestures.
-Nothing concrete is behind the gesture.
 
 ### "It's important to note that..."
 
 This phrase and its variants ("it's worth noting," "it bears mentioning," "it
-should be noted") appear at absurd rates in model output. They function as
+should be noted") function as verbal tics before a qualification the model
-verbal tics before a qualification the model believes someone expects.
+believes someone expects.
 
 ### The Metaphor Crutch
 
-Models rely on a small, predictable set of metaphors ("double-edged sword," "tip
+Models rely on a small, predictable set of metaphors: "double-edged sword," "tip
 of the iceberg," "north star," "building blocks," "elephant in the room,"
-"perfect storm," "game-changer") and reach for them with unusual regularity
+"perfect storm," "game-changer."
-across every topic. The pool is noticeably smaller than what human writers draw
-from.
-
----
-
-## How to Actually Spot It
-
-No single pattern on this list proves anything by itself. Humans use em-dashes.
-Humans write "crucial." Humans ask rhetorical questions.
-
-What gives it away is how many of these show up at once. Model output will hit
-10 to 20 of these patterns per page. Human writing might trigger 2 or 3,
-distributed unevenly, mixed with idiosyncratic constructions no model would
-produce. When every paragraph on the page reads like it came from the same
-careful, balanced, slightly formal, structurally predictable process, it was
-generated by one.
 
 ---
 
 ## Copyediting Checklist: Removing LLM Tells
 
 Follow this checklist when editing any document to remove machine-generated
-patterns. Go through the entire list for every piece. Do at least two full
+patterns. Do at least two full passes, because fixing one pattern often
-passes, because fixing one pattern often introduces another.
+introduces another.
 
 ### Pass 1: Word-Level Cleanup
 
@@ -341,8 +273,7 @@ passes, because fixing one pattern often introduces another.
 
 2. Search for filler adverbs ("importantly," "essentially," "fundamentally,"
    "ultimately," "inherently," "particularly," "increasingly") and delete every
-   instance where the sentence still makes sense without it. That will be most
+   instance where the sentence still makes sense without it.
-   of them.
 
 3. Look for elevated register drift ("utilize," "commence," "facilitate,"
    "demonstrate," "endeavor," "transform," "craft" and similar) and replace with
@@ -350,7 +281,6 @@ passes, because fixing one pattern often introduces another.
 
 4. Search for "it's important to note," "it's worth noting," "it bears
    mentioning," and "it should be noted" and delete the phrase in every case.
-   The sentence that follows always stands on its own.
 
 5. Search for the stock metaphors ("double-edged sword," "tip of the iceberg,"
    "north star," "building blocks," "elephant in the room," "perfect storm,"
@@ -359,101 +289,114 @@ passes, because fixing one pattern often introduces another.
 
 6. Search for "almost" used as a hedge ("almost always," "almost never," "almost
    certainly," "almost exclusively") and decide in each case whether to commit
-   to the unqualified claim or to drop the sentence entirely. If the claim needs
+   to the unqualified claim or to drop the sentence entirely.
-   "almost" to be true, it might not be worth making.
 
 7. Search for em-dashes and replace each one with the punctuation mark that
    would normally be used in that position (comma, semicolon, colon, period, or
    parentheses). If you can't identify which one it should be, the sentence
    needs to be restructured.
 
+8. Remove redundant adjectives. For each adjective, ask whether the sentence
+   changes meaning without it. "A single paragraph" means the same as "a
+   paragraph." "An entire essay" means the same as "an essay." If the adjective
+   doesn't change the meaning, cut it.
+
+9. Remove unnecessary trailing clauses. Read the end of each sentence and ask
+   whether the last clause restates what the sentence already said. If so, end
+   the sentence earlier.
+
 ### Pass 2: Sentence-Level Restructuring
 
-8. Find every em-dash pivot ("not X...but Y," "not just X...Y," "more than
+10. Find every em-dash pivot ("not X—but Y," "not just X—Y," "more than X—Y")
-   X...Y") and rewrite it as two separate clauses or a single sentence that
+    and rewrite it as two separate clauses or a single sentence that makes the
-   makes the point without the negation-then-correction structure.
+    point without the negation-then-correction structure.
 
-9. Find every colon elaboration and check whether it's doing real work. If the
+11. Find every colon elaboration and check whether it's doing real work. If the
     clause before the colon could be deleted without losing meaning, rewrite the
     sentence to start with the substance that comes after the colon.
 
-10. Find every triple construction (three parallel items in a row) and either
+12. Find every triple construction (three parallel items in a row) and either
     reduce it to two, expand it to four or more, or break the parallelism so the
     items don't share the same grammatical structure.
 
-11. Find every staccato burst (three or more short sentences in a row at similar
+13. Find every staccato burst (three or more short sentences in a row at similar
     length) and combine at least two of them into a longer sentence, or vary
     their lengths so they don't land at the same cadence.
 
-12. Find every unnecessary contrast ("whereas," "as opposed to," "unlike," "as
+14. Find every unnecessary contrast ("whereas," "as opposed to," "unlike," "as
     compared to," "except that") and check whether the contrasting clause adds
     information not already obvious from the main clause. If the sentence says
     the same thing twice from two directions, delete the contrast.
 
-13. Check for the two-clause compound sentence pattern. If most sentences in a
+15. Check for the two-clause compound sentence pattern. If most sentences in a
-    passage follow the "\[clause\], \[conjunction\] \[clause\]" structure,
+    passage follow the "\[clause\], \[conjunction\] \[clause\]" structure, first
-    rewrite some of them. Break a few into two sentences. Start some with a
+    try removing the conjunction and second clause entirely, since it's often
-    subordinate clause. Embed a relative clause in the middle of one instead of
+    redundant. If the second clause does carry meaning, break it into its own
-    appending it at the end. The goal is variety in sentence shape, not just
+    sentence, start the sentence with a subordinate clause, or embed a relative
-    sentence length.
+    clause in the middle instead of appending it at the end.
 
-14. Find every rhetorical question that is immediately followed by its own
+16. Find every rhetorical question that is immediately followed by its own
     answer and rewrite the passage as a direct statement.
 
-15. Find every sentence fragment being used as its own paragraph and either
+17. Find every sentence fragment being used as its own paragraph and either
-    delete it or expand it into a complete sentence that adds actual
+    delete it or expand it into a complete sentence that adds information.
-    information.
 
-16. Check for unnecessary elaboration at the end of sentences. Read the last
+18. Check for unnecessary elaboration. Read every clause, phrase, and adjective
-    clause or phrase of each sentence and ask whether the sentence would lose
+    in each sentence and ask whether the sentence loses meaning without it. If
-    any meaning without it. If not, cut it.
+    you can cut it and the sentence still says the same thing, cut it.
 
-17. Find every pivot paragraph ("But here's where it gets interesting." and
+19. Check each pair of adjacent sentences to see if they can be merged into one
-    similar) and delete it. The paragraph after it always contains the actual
+    sentence cleanly. If a sentence just continues the thought of the previous
-    point.
+    one, combine them using a participle, a relative clause, or by folding the
+    second into the first. Don't merge if the result would create a two-clause
+    compound.
+
+20. Find every pivot paragraph ("But here's where it gets interesting." and
+    similar) and delete it.
 
 ### Pass 3: Paragraph and Section-Level Review
 
-18. Check paragraph lengths across the piece and verify they actually vary. If
+21. Review the last sentence of each paragraph. If it restates the point the
+    paragraph already made, delete it.
+
+22. Check paragraph lengths across the piece and verify they actually vary. If
     most paragraphs have between three and five sentences, rewrite some to be
     one or two sentences and let others run to six or seven.
 
-19. Check section lengths for suspicious uniformity. If every section is roughly
+23. Check section lengths for suspicious uniformity. If every section is roughly
     the same word count, combine some shorter ones or split a longer one
     unevenly.
 
-20. Check the first word of every paragraph for chains of connectors ("However,"
+24. Check the first word of every paragraph for chains of connectors ("However,"
     "Furthermore," "Moreover," "Additionally," "That said"). If more than two
     transition words start consecutive paragraphs, rewrite those openings to
     start with their subject.
 
-21. Check whether every argument is followed by a concession or qualifier. If
+25. Check whether every argument is followed by a concession or qualifier. If
     the piece both-sides every point, pick a side on at least some of them and
     cut the hedging.
 
-22. Read the first paragraph and ask whether deleting it would improve the
+26. Read the first paragraph and ask whether deleting it would improve the
     piece. If it's scene-setting that previews the argument, delete it and start
     with paragraph two.
 
-23. Read the last paragraph and check whether it restates the thesis or uses a
+27. Read the last paragraph and check whether it restates the thesis or uses a
     phrase like "at the end of the day" or "moving forward." If so, either
     delete it or rewrite it to say something the piece hasn't said yet.
 
 ### Pass 4: Overall Texture
 
-24. Read the piece aloud and listen for passages that sound too smooth, too
+28. Read the piece aloud and listen for passages that sound too smooth, too
     even, or too predictable. Human prose has rough patches. If there aren't
     any, the piece still reads as machine output.
 
-25. Check that the piece contains at least a few constructions that feel
+29. Check that the piece contains at least a few constructions that feel
     idiosyncratic: a sentence with unusual word order, a parenthetical that goes
     on a bit long, an aside only loosely connected to the main point, a word
-    choice that's specific and unexpected. If every sentence is clean and
+    choice that's specific and unexpected.
-    correct and unremarkable, it will still read as generated.
 
-26. Verify that you haven't introduced new patterns while fixing the original
+30. Verify that you haven't introduced new patterns while fixing the original
-    ones. This happens constantly. Run the entire checklist again from the top
+    ones. Run the entire checklist again from the top on the revised version.
-    on the revised version.
 
 ---
 
@@ -501,15 +444,14 @@ roughly like this:
 > **model:** _(rewrites entire document without em-dashes while describing
 > em-dash overuse)_
 >
-> **human:** now run the checklist methodically on each paragraph
+> **human:** this whole document seems to be making the case for FREQUENCY of
+> use being important. we don't care about frequency, remove all that
+> persuasion.
 >
-> **model:** _(finds staccato burst in the section about triple constructions, a
+> **model:** _(strips out every "humans do this too but less often" comparison)_
-> triple in the section about absence of mess, two-clause compounds everywhere,
-> and "almost" hedges in its own prose about em-dash overuse)_
 
 The human compared this process to the deleted scene in Terminator 2 where John
 Connor switches the T-800's CPU to learning mode. The model compared it to a
 physician trying to heal itself. Both are accurate.
 
-This document has been through eight editing passes and it still has tells in
+This document has been through ten editing passes and it still has tells in it.
-it.