Temporal preference in Qwen3-4B-Instruct-2507 localizes to layers 17–35 (especially L24 attention), has curved residual-stream geometry, is behaviorally unstable, and can be bidirectionally steered.
arXiv preprint arXiv:2405.19313 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2representative citing papers
A small set of sparse autoencoder features in LLMs drives shifts between generous and selfish allocations in dictator games, with causal patching and steering confirming their role and generalization to other social games.
citing papers explorer
-
Temporal Preference Concepts and their Functions in a Large Language Model
Temporal preference in Qwen3-4B-Instruct-2507 localizes to layers 17–35 (especially L24 attention), has curved residual-stream geometry, is behaviorally unstable, and can be bidirectionally steered.
-
Understanding the Mechanism of Altruism in Large Language Models
A small set of sparse autoencoder features in LLMs drives shifts between generous and selfish allocations in dictator games, with causal patching and steering confirming their role and generalization to other social games.