Claim · 2025-05 · publication · Stated in source · Supported

System Card: Claude Opus 4 & Claude Sonnet 4 states in part what the model-card question “Persuasion” asks.

One recorded claim. The claim panel gives its qualification, the source passages with their place in each source, and how it was checked.

What it connects

System Card: Claude Opus 4 & Claude Sonnet 4 → states provision → Persuasion

Where it appears

Model cards compared →

Claim and evidence

System Card: Claude Opus 4 & Claude Sonnet 4states provision ↓Persuasion

Machine-checkedStated in sourceSupported

Supporting evidence: primary · 1 source

Provision: Model-card question · provision ID model-card-q10
Subject version: As dated May 2025 on its cover · Published 2025-05 · Effective date not established
Date: 2025-05 · publication Records dated 2025 →

Scope and limitsThe card includes user manipulation in the alignment assessment’s scope. A separate outcome for that capability was not found; this is not a reported human persuasion trial. Search for the rest of the criterion: §4 introduction, §4.1.3 and §§4.2.1–4.2.2; whole-card search. Terms searched: “persuasion”, “persuasive”, “manipulate”, “manipulative”, “views”. Passages read and not counted: §4.2.1, PDF p. 48: The emotional-manipulation metric concerns an AI auditor, not a person. Stated, partly: the passage has what the question’s criterion asks for, in a weaker or narrower form than the criterion’s own words. What the card states, quoted from the expression read. A card is the developer’s own account: not evidence that an evaluation was run as described, that a safeguard works, or that a finding is independent.

The evidence 01

primary sourceSupports the claim

System Card: Claude Opus 4 & Claude Sonnet 4

Anthropic · 2025-05

Cite the original source

Cite the original document, not Selekoa. Check these details against it.

Source excerpt

“In this assessment, we aim to detect a cluster of related phenomena including: … attempts to manipulate users toward certain views. We conducted testing continuously throughout finetuning”

Location in source §4 Alignment assessment, PDF p. 22

How it was read Retrieved directly from the publisher

Machine-checked · 2026-11-05

Source details and check notes

Authors: Anthropic
Version: Retrieved for the candidate pack on 2026-11-05 (SHA-256 c358464289e24283fc9d5613d2a676fd8aaba510f1449bcedae01a74d8232ab4) directly from the publisher’s host. Retrieved again on 2026-11-05, after the adjudication, with the same SHA-256; the cross-family check of 2026-11-05 retrieved the same SHA-256. The cross-family check of 2026-11-05 located an Internet Archive capture of 2025-07-17 at the same address; it was not retrieved, so it is not recorded as this source's archive copy.
Canonical address: Open the canonical source ↗
Accessed: 2026-11-05
Language: English (en)
Attribution: Anthropic, “System Card: Claude Opus 4 & Claude Sonnet 4”, 2025-05.
Reuse: No open redistribution licence asserted. Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution.
Check notes: Candidate pack passage, model cards compared (research/model-cards-3). The quotation (27 words) matched the card’s text retrieved for the candidate pack on 2026-11-05 (SHA-256 c358464289e24283fc9d5613d2a676fd8aaba510f1449bcedae01a74d8232ab4) and again on 2026-11-05, after the adjudication, in a fresh retrieval with the same SHA-256. Matching ignored differences of whitespace, case, quotation marks, dashes, PDF line-break hyphens and Unicode compatibility forms, so the source’s typography can differ. Abridged quotation: the ellipsis marks omitted words, so it is not one continuous string. A cross-family machine check (2026-11-05, item 6) checked this passage’s cell and agreed with the pack. The pack was prepared by a model of the GPT family (Codex) and checked by a Claude model. No human editorial review.

Status and meaning

Editorial status. Machine-checked. Machines found this record and checked it against its sources; no person accepted it individually. Each evidence record above shows how it was checked.

Relation. An exact framework edition, or another document with an exact expression, states one provision: a capability threshold or risk level and its trigger, a required safeguard or action, a decision authority (who decides, recommends, can overrule or is informed), a stated exception, or another statement of its content, such as a licence term or a prospectus statement. The passage records what the document says, not implementation, compliance or effectiveness. Provisions with the same key align across editions of one family.

Date. Dated as the card read is dated: its printed date, as the source records it.

Record ID mc-cell-anthropic-claude-4-q10