Claim and evidence Close ×
Meinke, Schoen, Scheurer, Balesni, Shah and Hobbhahn's paper of December 2024 states that large language models are no longer simple text generation systems but are deployed as agents capable of pursuing goals, and reports frontier models covertly pursuing goals given in context. Copy claim link
Copy claim reference
Machine-checked, not reviewed by an editor Stated in source Supported
Supporting evidence: primary · 1 source
Date 2024-12-06 · publication Records dated 2024 → Role Contesting Register (kind of text) Nonfiction Engagement Engages a premise. It answers the premise that language models only reproduce text by reporting models deployed as agents pursuing supplied goals. Scope and limits The goals in most evaluations were supplied in the prompt with an instruction to pursue them strongly; the paper reports capability for in-context scheming, not that models form goals of their own in ordinary use.
The evidence 02 primary source Supports the claim
Frontier Models are Capable of In-context Scheming ↗ arXiv preprint 2412.04984 (Apollo Research) · 2024-12-06
Copy source reference
Cite the original source Cite the original document, not Selekoa. Check these details against it.
Copy APA
Download BibTeX Download RIS Download CSL-JSON
Source excerpt “Large language models (LLMs) are no longer simple text generation systems but are increasingly trained and deployed as autonomous agents capable of independently pursuing goals and executing complex tasks.” Location in source Section 1, PDF p. 1
How it was read Retrieved directly from the publisher
Machine-checked · 2026-10-31
Source details and check notes Authors Alexander Meinke; Bronson Schoen; Jérémy Scheurer; Mikita Balesni; Rusheb Shah; Marius Hobbhahn Version Retrieved for the candidate pack on 2026-10-29 (SHA-256 075e622d98d6ee5504f9d60a8d89a9ccbc04648c498489a81d1d7274b8e0ab06). arXiv v1 of 6 December 2024. Canonical address Open the canonical source ↗ Accessed 2026-10-29 Language English (en) Attribution Alexander Meinke, Bronson Schoen, Jérémy Scheurer, Mikita Balesni, Rusheb Shah, Marius Hobbhahn, “Frontier Models are Capable of In-context Scheming”, 2024-12-06. 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, arguments about catastrophic risk from advanced AI (research/risk-debate). The quotation (29 words) matched the text retrieved for the candidate pack on 2026-10-29. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review. primary source Supports the claim
Frontier Models are Capable of In-context Scheming ↗ arXiv preprint 2412.04984 (Apollo Research) · 2024-12-06
Copy source reference
Cite the original source Cite the original document, not Selekoa. Check these details against it.
Copy APA
Download BibTeX Download RIS Download CSL-JSON
Source excerpt “making the potential of AI agents to engage in scheming behavior a concrete rather than theoretical concern” Location in source Abstract, PDF p. 1
How it was read Retrieved directly from the publisher
Machine-checked · 2026-10-31
Source details and check notes Authors Alexander Meinke; Bronson Schoen; Jérémy Scheurer; Mikita Balesni; Rusheb Shah; Marius Hobbhahn Version Retrieved for the candidate pack on 2026-10-29 (SHA-256 075e622d98d6ee5504f9d60a8d89a9ccbc04648c498489a81d1d7274b8e0ab06). arXiv v1 of 6 December 2024. Canonical address Open the canonical source ↗ Accessed 2026-10-29 Language English (en) Attribution Alexander Meinke, Bronson Schoen, Jérémy Scheurer, Mikita Balesni, Rusheb Shah, Marius Hobbhahn, “Frontier Models are Capable of In-context Scheming”, 2024-12-06. 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, arguments about catastrophic risk from advanced AI (research/risk-debate). The quotation (17 words) matched the text retrieved for the candidate pack on 2026-10-29. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review. Status and meaning Editorial status. Not reviewed by an editor . An exploratory record that no editor has accepted. This status does not say whether a person has read any linked passage. Each evidence record above shows how it was checked.
Relation. A passage of the document disputes the proposition or a premise it depends on. The record documents the text’s position, not who is right.
Date. Date of the arXiv v1 submission.
Record ID rd-r14-meinke-2024-in-context-scheming