TechniqueRLHF / Alignment8 recent entries4 Aug 2026RubricReviewer: From Direct Critique to Objective and Comprehensive Rubric-Driven Peer ReviewarXiv:2608.00005v1 Announce Type: new Abstract: Peer review at major venues is under unprecedented submission pressure, motivating the use of large language models (LLMs) as review assistants. Existin→5 Aug 2026Your Agentic LLMs Secretly Encode Latent Signals of Indirect Prompt-Injection ExposurearXiv:2608.02657v1 Announce Type: cross Abstract: Agentic LLMs are vulnerable to indirect prompt injection (IPI) attacks, e.g., malicious side-tasks hidden in external tool results. While many efforts
TechniqueRAG5 recent entries23 Jul 2026Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval ResultsarXiv:2607.20090v1 Announce Type: cross Abstract: Retrieval-augmented large language models frequently face contexts that interleave useful evidence with misleading statements or instruction-like cont→11 Aug 2026Defending Retrieval-Augmented Intrusion Detection Against Knowledge Poisoning and Prompt InjectionarXiv:2608.08100v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) enables large language models to classify network flows and generate human-readable incident reports by retrievin→14 Aug 2026Built a Mac app that actually uses Ollama for work, not just chat — with a calibration step for models with flaky tool callingMost Ollama frontends I've tried are chat UIs with tool calling bolted on as an afterthought, which is fine until you actually try to make a small local model do multi-step work and it silently emits →1 Sept 2026FocusAgent: Simple Yet Effective Ways of Trimming the Large Context of Web AgentsarXiv:2510.03204v2 Announce Type: replace Abstract: Web agents powered by large language models (LLMs) must process lengthy web page observations to complete user goals; these pages often exceed tens →7 Oct 2026RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG SystemsarXiv:2610.08571v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) systems are vulnerable to prompt-injection attacks embedded in retrieved content. We introduce RAG-PIBench, a ben
TechniqueAgents8 recent entries25 Sept 2026Decision Hijacking: Prompt Injection Attacks on Jev's Typed Probabilistic DecisionsarXiv:2609.28613v1 Announce Type: cross Abstract: Most studies of prompt injection focus on generative agents, leaving their effects on models with schema-defined outputs unclear. We examine these eff→2 Oct 2026Memetic Trojans: Social Contagions as Carriers of Adversarial Payloads in Agent NetworksarXiv:2610.00430v1 Announce Type: cross Abstract: Autonomous large language model (LLM) agents increasingly interact in network environments where adversarial content can propagate between agents. Kno→5 Oct 2026Prompted to Discriminate: Generalizing Malicious-Input Probes in the WildarXiv:2610.02413v1 Announce Type: new Abstract: LLM agents increasingly rely on activation probes as runtime monitors for prompt injection, jailbreaks, and unsafe requests, reading the model's own hid→5 Oct 2026Passing the Test You Trained On: Re-evaluating Prompt-Injection Detectors for LLM AgentsarXiv:2610.03448v1 Announce Type: cross Abstract: LLM agents increasingly screen tool outputs with small prompt-injection detectors, and teams choose among detectors by their scores on public benchmar→6 Oct 2026Self-Reflection Fine-Tuning: Enhancing Agent Security against Prompt Injection Attacks from Failure ExperiencearXiv:2610.04269v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly deployed in tool-augmented environments, but their reliance on external inputs makes them highly vuln→7 Oct 2026RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG SystemsarXiv:2610.08571v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) systems are vulnerable to prompt-injection attacks embedded in retrieved content. We introduce RAG-PIBench, a ben→9 Oct 2026One Word Opens the Gate: The Option-Channel Attack on Typed Decision Models as Agent GuardrailsarXiv:2610.12292v1 Announce Type: new Abstract: A typed decision model reads a piece of text and returns a probability over caller-defined options, each with a short written definition, generating no →12 Oct 2026LLMail-Inject: A Dataset from a Realistic Adaptive Prompt Injection ChallengearXiv:2506.09956v2 Announce Type: replace-cross Abstract: Indirect prompt injection attacks exploit a fundamental weakness of LLMs: executing unauthorized tasks from untrusted sources. Despite the pra
TechniqueFine-tuning8 recent entries29 May 2026Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral DetectionarXiv:2605.30189v1 Announce Type: cross Abstract: We show that LoRA adapters, the dominant distribution format for fine-tuned LLMs, can be reliably backdoored through training data poisoning while pre→9 Jun 2026Hiding in Plain Floats: Steganographic Carriers for Indirect Prompt and Content InjectionarXiv:2606.08403v1 Announce Type: cross Abstract: Text-centered prompt-injection defenses assume that the malicious signal is visible in one of the inspected text views. We study a reproducible LLM01-→11 Aug 2026ChronoState: Hidden Elapsed-Time Conditioning for Temporal-State Action Selection in Frozen-Backbone Language ModelsarXiv:2608.09124v1 Announce Type: new Abstract: Temporal decisions in language-model systems often depend on both symbolic task state and elapsed wall-clock time, such as cache expiration, job complet→2 Sept 2026HiveTraceGuard-Pro: A Compact Generative Guardrail for Prompt Injection, Jailbreaks, and Adversarial ObfuscationarXiv:2609.01046v1 Announce Type: cross Abstract: Production LLMs must handle inputs that attempt to override system instructions, bypass safety policies or elicit harmful responses. A common mitigati→15 Sept 2026PIDS-Bench: Evaluating Prompt-Injection Detectors Under Over-Defense, Obfuscation, and Distribution ShiftarXiv:2609.15017v1 Announce Type: cross Abstract: Prompt-injection detectors are typically evaluated using aggregate F1 on in-distribution test data, which offers limited insight into behavior under d→30 Sept 2026CALIBUDGET: Calibration-Guided Source Allocation for Fixed-Budget Mixed-Reasoning AdaptationarXiv:2609.36721v1 Announce Type: new Abstract: Fixed-budget adaptation from heterogeneous data sources requires deciding not only how much data to use, but how much exposure each source receives. Siz→5 Oct 2026SecJev: Bringing Security Expertise to System One Decision ModelsarXiv:2610.03073v1 Announce Type: cross Abstract: Security workflows need models that turn complex observations and explicit policies into decisions. System One models introduced by Jev return typed p→6 Oct 2026Self-Reflection Fine-Tuning: Enhancing Agent Security against Prompt Injection Attacks from Failure ExperiencearXiv:2610.04269v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly deployed in tool-augmented environments, but their reliance on external inputs makes them highly vuln
TechniqueMultimodal5 recent entries6 May 2026VisInject: Disruption != Injection -- A Dual-Dimension Evaluation of Universal Adversarial Attacks on Vision-Language ModelsarXiv:2605.01449v1 Announce Type: cross Abstract: Universal adversarial attacks on aligned multimodal large language models are increasingly reported with attack success rates in the 60-80% range, sug→28 May 2026MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated ContentarXiv:2605.28116v1 Announce Type: cross Abstract: Mobile graphical user interface (GUI) agents driven by vision-language models (VLMs) perceive the screen as rendered pixels and choose actions from wh→9 Sept 2026Beyond the Verdict: Evidence-Aligned Evaluation of Visual Prompt-Injection GuardrailsarXiv:2609.05535v1 Announce Type: new Abstract: Verdict-only evaluation does not reveal whether a vision-language model (VLM) used the visual evidence that should support its decision. We study this p→5 Oct 2026SecJev: Bringing Security Expertise to System One Decision ModelsarXiv:2610.03073v1 Announce Type: cross Abstract: Security workflows need models that turn complex observations and explicit policies into decisions. System One models introduced by Jev return typed p→5 Oct 2026Prompted to Discriminate: Generalizing Malicious-Input Probes in the WildarXiv:2610.02413v1 Announce Type: new Abstract: LLM agents increasingly rely on activation probes as runtime monitors for prompt injection, jailbreaks, and unsafe requests, reading the model's own hid
TechniqueSafety8 recent entries15 Sept 2026PIDS-Bench: Evaluating Prompt-Injection Detectors Under Over-Defense, Obfuscation, and Distribution ShiftarXiv:2609.15017v1 Announce Type: cross Abstract: Prompt-injection detectors are typically evaluated using aggregate F1 on in-distribution test data, which offers limited insight into behavior under d→23 Sept 2026ActGov: Governing LLM Agent Actions via Policy-Constrained ValidationarXiv:2609.24446v2 Announce Type: cross Abstract: Large language model (LLM) agents increasingly execute long-horizon workflows through external tools, allowing untrusted outputs to influence subseque→28 Sept 2026Monitor Jailbreaking: Evading Chain-of-Thought Monitoring Without Encoded ReasoningarXiv:2609.31121v1 Announce Type: new Abstract: Chain-of-thought (CoT) monitoring is a promising safety technique for reasoning models, enabling detection of problematic reasoning before models act. A→5 Oct 2026SecJev: Bringing Security Expertise to System One Decision ModelsarXiv:2610.03073v1 Announce Type: cross Abstract: Security workflows need models that turn complex observations and explicit policies into decisions. System One models introduced by Jev return typed p→5 Oct 2026Prompted to Discriminate: Generalizing Malicious-Input Probes in the WildarXiv:2610.02413v1 Announce Type: new Abstract: LLM agents increasingly rely on activation probes as runtime monitors for prompt injection, jailbreaks, and unsafe requests, reading the model's own hid→6 Oct 2026Self-Reflection Fine-Tuning: Enhancing Agent Security against Prompt Injection Attacks from Failure ExperiencearXiv:2610.04269v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly deployed in tool-augmented environments, but their reliance on external inputs makes them highly vuln→7 Oct 2026RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG SystemsarXiv:2610.08571v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) systems are vulnerable to prompt-injection attacks embedded in retrieved content. We introduce RAG-PIBench, a ben→9 Oct 2026One Word Opens the Gate: The Option-Channel Attack on Typed Decision Models as Agent GuardrailsarXiv:2610.12292v1 Announce Type: new Abstract: A typed decision model reads a piece of text and returns a probability over caller-defined options, each with a short written definition, generating no