Frontier Models are Capable of In-context Scheming · HackerPapers discussions

Frontier Models are Capable of In-context Scheming — HN submission

Tue, 30 Dec 2025 20:26:54 GMT

Frontier Models are Capable of In-context Scheming

2 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

Frontier Models are Capable of In-context Scheming — HN submission

Tue, 27 May 2025 18:33:42 GMT

Frontier Models are Capable of In-context Scheming

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Frontier Models are Capable of In-context Scheming — HN submission

Tue, 20 May 2025 22:11:18 GMT

Frontier Models are Capable of In-context Scheming

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Frontier Models are Capable of In-context Scheming — HN submission

Mon, 13 Jan 2025 22:43:34 GMT

Frontier Models are Capable of In-context Scheming

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Frontier Models are Capable of In-context Scheming — HN submission

Mon, 06 Jan 2025 01:56:41 GMT

Frontier Models are Capable of In-context Scheming

4 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

Frontier Models are Capable of In-context Scheming — mentioned by abrichr

Fri, 13 Dec 2024 20:07:05 GMT

https://arxiv.org/abs/2412.04984 > Our findings demonstrate that frontier models now possess capabilities for basic in-context scheming [covertly pursuing misaligned goals], making the potential of AI agents to engage in scheming behavior a concrete rather than theoretical concern.

Paper and discussion history. A mention is not an endorsement.

Frontier Models are Capable of In-context Scheming — HN submission

Thu, 12 Dec 2024 22:31:30 GMT

Frontier Models are Capable of In-context Scheming

10 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.