Stochastic CHAOS: Why Deterministic Inference Kills, and Distributional Variability Is the Heartbeat of Artifical Cognition

arXiv:2601.07239 · cs.AI · Submitted 2026-01-12 · Read on arXiv

cs.AI

Submitted: 2026-01-12

Updated: 2026-10-11

Code: https://github.com/yale-nlp/InstruSum

Terminology

• Accuracy — 1× in this paper · explained in 1 episode
• Alignment — 1× in this paper · explained in 1 episode

Sources

• Non-Determinism of "Deterministic" LLM Settings
• Red Teaming Language Models with Language Models
• Alignment faking in large language models
• The Curious Case of Neural Text Degeneration
• Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
• Existence and uniqueness of solutions in the Lipschitz space of a functional equation and its application to the behavior of the paradise fish
• Large Language Models are Zero-Shot Reasoners
• Deterministic Implementations for Reproducibility in Deep Reinforcement Learning
• Abstractive Text Summarization Using Sequence-to-Sequence RNNs and Beyond
• Offline RL for Natural Language Generation with Implicit Language Q Learning
• Are Emergent Abilities of Large Language Models a Mirage?
• GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
• Self-Consistency Improves Chain of Thought Reasoning in Language Models
• Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
• Tree of Thoughts: Deliberate Problem Solving with Large Language Models
• Deterministic Inference across Tensor Parallel Sizes That Eliminates Training-Inference Mismatch
• SGCCNet: Single-Stage 3D Object Detector With Saliency-Guided Data Augmentation and Confidence Correction Mechanism

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems