← Back to all news
Hugging Face Trending Papers

An Omitted Mode Is a Rare Rule: The Sampling-Verification Danger Law in Continuous Code World Models

The paper investigates the safety of Code World Models, where a language model generates an executable world model that a planner uses, and the model is accepted if it reproduces sampled transitions. It defines the pipeline’s danger as the expected risk, showing that the probability of missing a critical event across N independent rollouts is (1‑r)^N, and that an additional acceptance sample adds to the exponent. Experiments on hybrid instruments reveal that mode‑blind models can be exploited, and the authors provide theoretical bounds on localization budgets and demonstrate that acceptance only guarantees sample consistency, covering about two percent of the planner’s queries.

Related stories

arXiv AI
Aug 19

An Omitted Mode Is a Rare Rule: The Sampling-Verification Danger Law in Continuous Code World Models

The paper investigates the safety of Code World Models, where a language model generates executable world models that a planner uses. It shows that accepting a model based on sampled transitions only guarantees sample consistency, not full safety, because the probability of missing critical events decays as (1‑r)^N. Experiments on hybrid instruments reveal that omitted mode‑boundaries can severely limit planner performance, and that even sophisticated LLMs (GPT‑5.x) struggle to repair such omissions in higher‑dimensional settings.

By Javier Aguilar Mart\'in
llms
More like this →
arXiv Machine Learning
Aug 31

An Enclosed Mode Is a Gauge Choice: Topology Relative to Reach in Certified Code World Models

The paper studies how a code‑world model can be perfectly accurate on the portion of the state space that a sampling gate can observe while potentially being arbitrarily wrong elsewhere. By treating the unobservable interior as an annular freeze mode, the authors formalize the notion of reach and show that acceptance with certainty fixes the model only on the reachable query set, leaving the rest as a gauge. Experiments on a minimal ring instrument demonstrate that a single channel width parameter can move the model through regimes of being unfalsifiable and harmless, falsifiable and costly, or instantly falsified, illustrating how topology relative to reach governs danger, repair, and mitigation strategies.

By Javier Aguilar Mart\'in
llms
More like this →
arXiv AI
Sep 25

Stale Does Not Mean Unsafe: Guard Precision for Tool-Using LLM Agents under Infrastructure State Races

The paper investigates how tool‑using language‑model agents can safely commit changes to infrastructure when external state may change between read and commit. By distinguishing invalidating races from predicate‑preserving and irrelevant ones, the authors evaluate three commit‑time guard granularities—global epoch, read‑set version, and semantic commit predicate—using a deterministic simulator and three quantized model families. The study finds that only the complete predicate guard consistently eliminates unsafe commits, while freshness‑based guards block a large proportion of benign races and model‑side signals fail to replace precise semantic enforcement.

By Zihao Zheng, Jiayu Long, Baichuan Li, Junyi Yao
llmsagentsefficiencysafety
More like this →
arXiv AI
Sep 7

SiLR: Structure-Preserving Admission and Process Reward for LLM Tool Agents

SiLR introduces a structure‑preserving admission and process reward mechanism for large language model (LLM) tool agents. Unlike traditional scalar‑score gates that can trap agents in plateau trajectories, SiLR shadow‑executes each proposal and admits it based on a product order over branch‑level violation states, ensuring safe and recoverable actions. Experiments on Gym‑ANM and CityLearn benchmarks show SiLR consistently recovers all multi‑action episodes and outperforms scalar gates, while also providing a robust reward signal for policy learning.

By Chenyu Zhou, Qiliang Jiang, Shuning Wu, Xu Zhou
llmsagentsbenchmarks
More like this →