Financial Context-Rich Strategy for Actor-Critic based Agents: A Deep Reinforcement Learning Approach for Stock Trading
Resumo indisponível. Consulte o paper original.
PAPER / ARXIV:2610.07935
Jingyu Zhang, Shruti Palaskar, Daniel Khashabi, Benjamin Van Durme, Leon A. Gatys, Joseph Yitan Cheng
RESUMO
LLM agents are increasingly capable of executing complex tasks and of recursively improving themselves on easy-to-verify objectives. Since alignment is much harder to verify, this creates a growing risk of capabilities increasing without appropriate safety alignment. We propose SIGMA, a data generation and training pipeline enabling alignment self-improvement that generalizes to out-of-distribution settings. Given only a Model Spec stating the model's desired behavior, SIGMA leverages a model's reasoning capabilities to strengthen its own safety reasoning via spec-guided task synthesis and self-judged alignment training. Despite training only on single-turn chat data, SIGMA improves safety alignment in multi-turn agentic environments (AgentHarm harmfulness decreases from 22.6 to 14.8; Agentic Misalignment decreases from 79.1 to 3.8), outperforming Deliberative Alignment and Constitutional AI baselines while retaining general capability.
NO MESMO MAPA
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.