calmkeepai

3 ms·

1.
by calmkeepai 6mo ago
That’s not what it does- while some details remain proprietary I can say that it’s not prompt injection it’s sequence management, much more akin to the logic of retaining form over arcs of change than any kind of prompting method
2.
LLM Drift in Long Sessions: Claude 60% vs. 85% Integrity After 25 Turns(calmkeep.ai)
2 points by calmkeepai 6mo ago|1 comments
3.
LLM drift-Claude vs. Calmkeep: 25-turn Code (60% vs. 85%) & Legal (50% vs. 100%)(calmkeep.ai)
2 points by calmkeepai 6mo ago|0 comments
4.
Project Itohs Harmony and the under explored extremes of alignment theory
1 points by calmkeepai 7mo ago|0 comments
5.
by calmkeepai 7mo ago
I’ve found an interesting model to think about to be production crews similar to in the television world but potentially be something worth using as one’s mental model of how agents and people working alongside agents should coordinate, ra
6.
by calmkeepai 7mo ago
It’s crazy how if these came out 16 years ago they would be dominant in ad campaigns and marketing but today , while the remain aspirationally pervasive, Apple just doesn’t saturate their marketing. This speaks to some larger changes in the
7.
by calmkeepai 7mo ago
The immersion into the time and place was fantastic, the surreal elements being bold , outlandish, and unexpected were great. The time jump at the end was interesting. a great piece of work that some felt divided over as a general audience
8.
LLM drift-Claude vs. Calmkeep: 25-turn Code (60% vs. 85%) & Legal (50% vs. 100%)(calmkeep.ai)
2 points by calmkeepai 7mo ago|0 comments