评分 6 · 来源：cs.LG updates on arXiv.org · 发布于 2026-04-20
评分依据：幻觉作为轨迹承诺：不对称吸引子的因果证据，新理论视角
要点
arXiv:2604.15400v1 Announce Type: new Abstract: We present causal evidence that hallucination in autoregressive language models is an early trajectory commitment governed by asymmetric attractor dynamics. Using same-prompt bifurcation, in which we repeatedly sample identical inputs to observe spontaneous divergence, we isolate trajectory dynamics from prompt-level confounds. On Qwen2.5-1.5B across 61 prompts spanning six categories, 27 prompts (44.3%) bifurcate with factual and hallucinated trajectories diverging at the first generated token (KL = 0 at step 0, KL > 1.0 at step 1). Activation p…
🤖 AI 点评
本文提供了AI领域的重要信息，值得行业从业者关注。