AI and model evaluation
Amanda Askell
What changes in a judgment of sincerity when only the source label changes?
Read the letterbeforeword · Letters published by their author
What in a text supports a claim that someone understands, feels pain or knows something?
Letters addressing specific wording in articles, books and talks. English texts, Russian translations and the available correspondence record.
The letters ask what supports particular steps: from identical words to a judgment of sincerity; from a prediction result to understanding; from a model description to an attribution of pain.
A public record lets readers compare the wording, beforeword’s question and any published reply. Definitions, criteria and explanations—including those proposed by beforeword—are also part of the examination.
The sentence stays the same. The source label changes. What grounds support a judgment of sincerity in each case?
Read the letterA starting point
AI and model evaluation
What changes in a judgment of sincerity when only the source label changes?
Read the letterLanguage and reading
What is added to the wording when inferring what the writer meant?
Read the letterConsciousness and experience
What supports the move from described behavior to an attribution of experienced pain?
Read the letterWhat changes in a judgment of sincerity when only the source label changes?
What supports each claim in an explanation of a model’s behavior?
What supports different readings of “I believe this” from a human and an AI?
What supports reading a written sentence as a report about consciousness?
What supports admitting a written record as an observation for AI training?
What supports saying that two formulations “say the same thing”?
What supports reading successful prediction as evidence of understanding?
What supports reading action-recognition results as progress toward understanding the world?
What justifies reading model output as an account of its reasoning?
What supports reading an experimental report as evidence for a causal assumption?
What supports treating two instances of a pattern as the same self?
What supports reading a report as evidence of “pure awareness”?
What supports reading a sentence as an explanation of perception?
What supports reading a passage as evidence of a human preference?
What supports attributing the same idea to two people?
What supports treating a writer as committed to justifying a claim?
What supports attributing mutual knowledge in a shared viewing task?
What distinguishes changing the words from changing the explanation?
What warrants calling a revision of text a correction of thought?
What is added to the wording when inferring what the writer meant?
What supports saying that a reader knows a claim is true?
What supports describing tissue’s bioelectrical state as memory of the body’s shape?
What supports the move from described behavior to an attribution of experienced pain?
What claim favors a consciousness interpretation over an alternative explanation?
No matches. Try another search or reset the filters.
The author is publishing sent letters about specific wording in articles, books and talks. Publication does not imply that a recipient participates in the project or agrees with beforeword.
The English texts are transcribed from screenshots supplied by the author. Screen line breaks are joined into paragraphs, and overlapping passages between images are included once. Wording and punctuation are retained; Russian translations and beforeword introductions are labeled separately.
Dates and addresses are limited to what the supplied materials show. Where a header shows only “Friday” or hides an additional recipient, the letter page says so. Sending, receipt, forwarding and a substantive reply are different records.
No replies have been published in this collection yet. This status does not mean agreement, rejection or an inability to object. beforeword’s own wording and proposed grounds remain open to the same examination.