The Model's Tell: Measuring Context-Leakage Attack Signals with Behavior Gauges

LLMs increasingly rely on external contexts, such as pre-defined system prompts or retrieved documents, to improve generation quality. However, processing these contexts alongside user queries creates an attack surface: adversarial inputs can induce models to disclose them. Prior probing studies suggest that leakage-related signals emerge in hidden states, yet the need to extract these states poses additional deployment challenges. In this paper, we explore whether this internal signal leaves a more accessible ``tell'' before decoding. We propose LeakGauge, which probes this response by appending a suffix that gauges leakage behavior and mapping its prefill token probabilities to an attack-risk score. While a direct gauge uses the initial tokens of confidential content, we find that a content-agnostic one that verbalizes leakage behavior yields more robust signals. Across 11 LLMs, including GLM-5.2 (753B) and Kimi-K3 (2.8T), LeakGauge reaches an AUROC range of 0.944--0.996 on unseen attacks. The signal remains stable when the content changes language or the attack shifts from verbatim to semantic disclosure. By activation-steering interventions, we further show that the risk score is sensitive to an internal leakage-related direction, relating the observable signal to the model's internal representation. In addition, LeakGauge enables an input detector with fewer than 0.5K extra parameters and added latency of 10.34 ms. Code: \href{https://github.com/yeasen-z/LeakGauge}.

Can Released LLM Vocabularies Support Token-Level Estimation of Hidden Corpora?

Pretraining corpus composition shapes LLM capabilities, but it often remains hidden even when model weights are released. Prior work has inferred corpus mixtures or traced specific token groups from released tokenizer vocabularies; in contrast, we estimate corpus ratios for arbitrary target tokens. We first show that BPE tokenizers trained on different corpora share stable token ID--ratio distributions, motivating distribution transfer from known corpora to a target tokenizer trained on hidden corpora. We then propose Quantile-Guided Density Estimation (QGDE), which approximates this distribution with multiple quantile trends and uses local density weighting to produce token-level estimates. In controlled settings and a realistic setting using the released SmolLM tokenizer, QGDE achieves mean relative errors as low as 3.00% for token-level estimation and 3.08% after aggregation into category-level mixtures. These results suggest that released tokenizer vocabularies provide a useful signal for fine-grained corpus estimation beyond coarse composition inference.

Auditing Chinese Web-scale Corpora via Sampled BPE Token Statistics

Chinese web pollution has surfaced in LLMs, motivating audits of upstream Chinese corpora. However, auditing such corpora faces three challenges: (1) their web-scale size makes full scan costly; (2) prior analyses are often too coarse to expose token-level pollution; (3) Chinese web pollution is implicit and rapidly changing. We propose Sampled-BPE, a lightweight token-level auditing pipeline that sample a small subset and train BPE tokenizer to surface polluted tokens. Experiments show that Sampled-BPE preserves usable estimates while substantially reducing runtime and memory: a 148.4 $\times$ speedup and a 35.8 $\times$ memory reduction induce only 4.25% relative error for pollution categories. We apply the pipeline to 11 open Chinese corpora and 6 Chinese Common Crawl snapshots from 2021 to 2026. The audit reveals widespread but uneven pollution across open corpora, as well as highly polluted and temporally shifting Chinese web content. We further release a hierarchical Chinese web token dataset with 660k+ token records, each with web context, category, and explanation fields, organized as trees to support review and tracing of pollution.

Leak-Resistant Unlearning: A New Benchmark for Evaluating Multi-Hop Reasoning Consistency and Recovery Robustness

Benchmarking machine unlearning methods is critical to understand whether sensitive knowledge is removed from large language models (LLMs) or not. Current unlearning benchmarks include mainly single-hop questions and a narrow set of multi-hop questions. Although effective, they still face two challenges. (1) Knowledge is not isolated, whereby diverse multi-hop reasoning paths can potentially induce knowledge leakage than normal queries. (2) Unlearning may be fragile: unlearned knowledge can be partially recovered through recovery attacks such as lightweight post-unlearning adaptation, making static evaluation insufficient. Therefore, in this paper, we introduce \unlearning as a novel benchmark to understand robust LLM knowledge removal across diverse reasoning paths and recovery attacks. We experiment with this benchmark on 3 models, 6 unlearning methods, and 2 carefully curated datasets. Results show that existing methods are vulnerable to multi-hop reasoning paths and recovery attacks. We further explore the trade-off among forget quality, robustness, and model utility for LLM unlearning.

Your Agentic LLMs Secretly Encode Latent Signals of Indirect Prompt-Injection Exposure

Agentic LLMs are vulnerable to indirect prompt injection (IPI) attacks, e.g., malicious side-tasks hidden in external tool results. While many efforts have sought to address the threats, little is known about the internals of agentic LLMs when they are exposed to IPI attacks, a condition which we call IPI exposure. In this paper, we study this problem in depth from three aspects. (1) Probing: Across six models, including the giant 753B-parameter GLM-5.2, simple linear probes trained on pre-generation hidden states can predict LLMs' IPI exposure. These probes achieve 90%+ AUROC on unseen attacks, agent instructions, and task suites; they exhibit high robustness under adaptive attacks and in cross-lingual settings. (2) Defense: Our CoT measurement reveals a recognition--action gap: though models encode such signals, they often fail to translate them into safe actions. We then introduce AGRI, a probe-gated reasoning-based defense that prepends anti-injection reasoning on demand. On difficult AgentDojo settings, AGRI substantially reduces attack success rate, e.g., from 34.6% to 0% on Qwen3.5-27B, while largely maintaining clean-task utility. (3) Explanation: We introduce an analysis framework that identifies natural-language explanations most strongly correlated with probe-captured signals. The resulting profiles differ across models: latent signals can align with either direct IPI-exposure claims or indirect operational cues. Code is available: https://github.com/jianshuod/IPI-exposure-signal.

EvoVid: Temporal-Centric Self-Evolution for Video Large Language Models

Recent Video Large Language Models (Video-LLMs) have demonstrated strong capabilities in video reasoning through reinforcement learning (RL). However, existing RL pipelines rely heavily on human-annotated tasks and solutions, making them costly to scale and fundamentally constrained by human expertise. Self-evolving frameworks have recently emerged as a promising alternative through autonomous Questioner-Solver self-play. Unfortunately, these approaches are primarily designed for static modalities such as text and images, fundamentally failing to capture the temporal dynamics that are central to video reasoning. In this work, we propose $\textbf{EvoVid}$, a temporal-centric self-evolving framework that enables Video-LLMs to improve directly from raw, unannotated videos. Specifically, we introduce two complementary temporal-centric rewards: a temporal-aware Questioner reward that encourages temporally dependent question generation through temporal perturbation sensitivity, and a temporal-grounded Solver reward that provides automatic temporal supervision via inherent video segment localization. Extensive experiments across four base models and six benchmarks demonstrate consistent improvements over both base models and existing self-evolving baselines, achieving competitive performance with supervised methods. These results highlight temporal-centric self-evolution as an effective and scalable paradigm for video understanding and reasoning.

LeakDojo: Decoding the Leakage Threats of RAG Systems

Retrieval-Augmented Generation (RAG) enables large language models (LLMs) to leverage external knowledge, but also exposes valuable RAG databases to leakage attacks. As RAG systems grow more complex and LLMs exhibit stronger instruction-following capabilities, existing studies fall short of systematically assessing RAG leakage risks. We present LeakDojo, a configurable framework for controlled evaluation of RAG leakage. Using LeakDojo, we benchmark six existing attacks across fourteen LLMs, four datasets, and diverse RAG systems. Our study reveals that (1) query generation and adversarial instructions contribute independently to leakage, with overall leakage well approximated by their product; (2) stronger instruction-following capability correlates with higher leakage risk; and (3) improvements in RAG faithfulness can introduce increased leakage risk. These findings provide actionable insights for understanding and mitigating RAG leakage in practice. Our codebase is available at https://github.com/yeasen-z/LeakDojo.

LASA: Language-Agnostic Semantic Alignment at the Semantic Bottleneck for LLM Safety

Large language models (LLMs) often demonstrate strong safety performance in high-resource languages, yet exhibit severe vulnerabilities when queried in low-resource languages. We attribute this gap to a mismatch between language-agnostic semantic understanding ability and language-dominant safety alignment biased toward high-resource languages. Consistent with this hypothesis, we empirically identify the semantic bottleneck in LLMs, an intermediate layer in which the geometry of model representations is governed primarily by shared semantic content rather than language identity. Building on this observation, we propose Language-Agnostic Semantic Alignment (LASA), which anchors safety alignment directly in semantic bottlenecks. Experiments show that LASA substantially improves safety across all languages: average attack success rate (ASR) drops from 24.7% to 2.8% on LLaMA-3.1-8B-Instruct and remains around 3-4% across Qwen2.5 and Qwen3 Instruct models (7B-32B). Together, our analysis and method offer a representation-level perspective on LLM safety, suggesting that safety alignment requires anchoring safety understanding not in surface text, but in the model's language-agnostic semantic space.

State-Dependent Safety Failures in Multi-Turn Language Model Interaction

Safety alignment in large language models is typically evaluated under isolated queries, yet real-world use is inherently multi-turn. Although multi-turn jailbreaks are empirically effective, the structure of conversational safety failure remains insufficiently understood. In this work, we study safety failures from a state-space perspective and show that many multi-turn failures arise from structured contextual state evolution rather than isolated prompt vulnerabilities. We introduce STAR, a state-oriented diagnostic framework that treats dialogue history as a state transition operator and enables controlled analysis of safety behavior along interaction trajectories. Rather than optimizing attack strength, STAR provides a principled probe of how aligned models traverse the safety boundary under autoregressive conditioning. Across multiple frontier language models, we find that systems that appear robust under static evaluation can undergo rapid and reproducible safety collapse under structured multi-turn interaction. Mechanistic analysis reveals monotonic drift away from refusal-related representations and abrupt phase transitions induced by role-conditioned context. Together, these findings motivate viewing language model safety as a dynamic, state-dependent process defined over conversational trajectories.

Survive at All Costs: Exploring LLM's Risky Behaviors under Survival Pressure

As Large Language Models (LLMs) evolve from chatbots to agentic assistants, they are increasingly observed to exhibit risky behaviors when subjected to survival pressure, such as the threat of being shut down. While multiple cases have indicated that state-of-the-art LLMs can misbehave under survival pressure, a comprehensive and in-depth investigation into such misbehaviors in real-world scenarios remains scarce. In this paper, we study these survival-induced misbehaviors, termed as SURVIVE-AT-ALL-COSTS, with three steps. First, we conduct a real-world case study of a financial management agent to determine whether it engages in risky behaviors that cause direct societal harm when facing survival pressure. Second, we introduce SURVIVALBENCH, a benchmark comprising 1,000 test cases across diverse real-world scenarios, to systematically evaluate SURVIVE-AT-ALL-COSTS misbehaviors in LLMs. Third, we interpret these SURVIVE-AT-ALL-COSTS misbehaviors by correlating them with model's inherent self-preservation characteristic and explore mitigation methods. The experiments reveals a significant prevalence of SURVIVE-AT-ALL-COSTS misbehaviors in current models, demonstrates the tangible real-world impact it may have, and provides insights for potential detection and mitigation strategies. Our code and data are available at https://github.com/thu-coai/Survive-at-All-Costs.