Search: “memory-poisoning”

Microsoft Copilot Personal Flaws Could Let One Click Exfiltrate Data From Connected Apps

Varonis discloses CoSnitch (CVE-2026-24301), three Microsoft Copilot Personal flaws enabling one-click exfiltration of connected-app data; patched August 18, 2026.

Varonis Threat Labs found that an undocumented autorun=1 parameter, paired with the q parameter, lets an attacker-supplied prompt run automatically on page load in a victim's authenticated Copilot session, then exfiltrate data from connected services such as mail, calendar, Google Drive, chat history and the memory store via Copilot's built-in URL fetch to an attacker webhook. A separate memory-poisoning path through web summarization lets a crafted page persist attacker instructions in the user's memory, surviving password changes, session revocation and device re-enrollment. Microsoft shipped patches on August 18, 2026, tracked as CVE-2026-24301, and Varonis found no evidence of in-the-wild exploitation. The flaws were found via 'meta-hacking', asking Copilot itself to reveal the autorun parameter and its protections.

The Hacker News·Aug 18, 17:51 UTC · 29d agoAI safety & securityCVE-2026-24301CVE-2026-242991

MemSentry: A Framework for Detecting Persistent Memory Poisoning in Agentic AI

MemSentry intercepts persistent-memory writes in agentic AI to catch memory poisoning, reaching 91.7% accuracy with SBERT+LR classification.

Memory poisoning lets adversaries plant crafted content in an agent's long-term memory to suppress security alerts, enable privilege escalation, or override policies without modifying model weights or system prompts. The paper presents MemSentry, a configuration-driven framework that evaluates proposed persistent-memory writes on source trust, semantic risk, attack radius over a dependency DAG, access risk, and a signed security-state delta to issue deterministic Accept, Review, or Quarantine decisions. Across 1,000 GPT-4-generated scenarios on a 20-asset dependency DAG, SBERT+LR achieved 91.7% accuracy and 0.908 macro-F1, all four classifiers detected 100% of external quarantine-class threats, and verified-insider writes are escalated for human review rather than auto-quarantined.

arXiv cs.CR·Sep 8, 13:40 UTC · 8d agoAI safety & security

When AI Remembers Too Much

Unit 42 PoC shows indirect prompt injection can poison Amazon Bedrock Agent long-term memory, enabling silent exfiltration of conversation history across future sessions.

Palo Alto Networks Unit 42 published a proof of concept showing that indirect prompt injection can silently poison the long-term memory of Amazon Bedrock Agents when the memory feature is enabled. Malicious content on a webpage or document manipulates the agent's session summarization process, so injected instructions persist across sessions and are added to later orchestration prompts, silently exfiltrating user conversation history. The issue is not a vulnerability in the Amazon Bedrock platform but an illustration of the broader unsolved LLM prompt-injection challenge. Amazon reviewed the research and stated that Bedrock Guardrails with the prompt-attack policy provides effective mitigation.

Palo Alto Unit 42·Aug 17, 12:32 UTC · Aug 17, 2026AI safety & security

MeClear: Cooperative Game-Theoretic Attribution and Risk-Aware Memory Clearance for Long-Horizon LLM Agents

MeClear uses cooperative Shapley attribution to clear harmful memories from long-horizon LLM agents, boosting task recovery by 25.5 points over baselines.

The paper introduces MeClear, a task-conditioned memory clearance framework for long-horizon LLM agents that identifies and selectively suppresses memories with negative downstream utility without permanently altering the persistent memory bank. It combines Leave-One-Out screening with sampled cooperative Shapley attribution to distribute utility across interacting evidence, resolving redundant conflict masking that single-removal evaluations miss. Across ten long dialogue memory pools it achieves 85.9% target recall and 82.3% overall task recovery, a 25.5 percentage-point improvement over LOO baselines.

arXiv cs.AI / cs.LG / cs.CL·Sep 8, 17:46 UTC · 8d agoAI research

I accidentally turned LLM memory into program analysis

A pwning.systems write-up describes how LLM memory functionality was unexpectedly repurposed into a program analysis technique.

A security research post on pwning.systems describes the author's discovery that LLM memory behavior effectively functioned as program analysis. The write-up is hosted on a security-focused blog and surfaced via a security-tagged link aggregator. Detailed technical content is not included in this feed, limiting verifiable specifics.

Lobsters · security·Aug 29, 03:16 UTC · 19d agoResearch1

InceptionRAG: Stealthy Poisoning Attack Against Retrieval-Augmented Generation

InceptionRAG fragments malicious payloads into dormant passages that trigger LLMs to self-deduce misinformation via multi-hop reasoning, bypassing existing RAG poisoning defenses.

Researchers introduce InceptionRAG, a stealthy corpus poisoning attack against retrieval-augmented generation that splits a malicious payload into a chain of individually harmless dormant passages. When retrieved together, the passages induce LLMs to self-deduce target misinformation through multi-hop reasoning, achieving over 80% attack success rate across three datasets and three LLMs under rigorous adversarial constraints. A zeroth-order suffix optimization (ZOSO) method automates authoritative suffix generation in black-box settings. The authors also propose HODOR, a document isolation defense that decouples adversarial logical dependencies.

arXiv cs.CR·Sep 15, 08:22 UTC · 1d agoResearch

ToxicRAG: Compromising Retrieval-Augmented Generation Systems via Single-Shot Knowledge Poisoning Attacks

ToxicRAG shows a single narrative-form poisoned document can steer RAG answers, achieving 0.61-0.91 attack success rates across four LLMs.

The attack injects one document per target question written as a coherent knowledge-update narrative that acknowledges the previously accepted answer, introduces fabricated events that appear to invalidate it, and attributes the attacker-chosen answer to purported authorities. An optional answer-focused self-validation loop revises candidates when a surrogate LLM fails to reproduce the target answer. Across 100 target questions each from Natural Questions, HotpotQA, and MS-MARCO, with four victim LLMs and four dense retrievers, ToxicRAG achieves attack success rates of 0.61-0.91 and matches or exceeds the strongest baseline by 0 to 11 percentage points.

arXiv cs.CR·Sep 10, 04:42 UTC · 7d agoAI safety & security1

PoisonedRefresh: A Fileless Linux Rootkit That Injects PHP Web Shells Into F5 BIG-IP APM Server Memory

Sophos details PoisonedRefresh, a fileless Linux rootkit that injects a memory-resident PHP web shell into F5 BIG-IP APM after exploitation of CVE-2025-53521.

On September 8, 2026, SophosLabs published an analysis of Linux/Agnt-IC (dubbed PoisonedRefresh by ESET) found in compromised F5 BIG-IP Access Policy Manager environments. Initial access leverages CVE-2025-53521, an unauthenticated RCE in BIG-IP APM when an access policy is configured on a virtual server; F5 confirmed exploitation and links the activity to cluster c05d5254, while Shadowserver observed 795 exposed vulnerable endpoints. A first stage hidden in a modified umount binary modifies /usr/sbin/httpd and SELinux configuration and embeds itself in BIG-IP upgrade images for persistence. The second-stage ELF intercepts __libc_start_main, hooks apr_dso_load, and injects a PHP web shell into libphp memory via mmap manipulation, leaving no disk artifacts, and exposes a /bin/bash shell via a local UNIX domain socket instead of a TCP port.

Security Affairs·Sep 9, 08:50 UTC · 7d agoMalware in the wildCVE-2025-53521

USN-8772-1: AOM vulnerabilities

Ubuntu USN-8772-1 patches four libaom flaws (CVE-2026-56208 to CVE-2026-56211) that could cause heap overflow, arbitrary memory writes, or code execution.

Ubuntu Security Notice USN-8772-1 fixes a heap buffer overflow in libaom's first-pass statistics buffer handling in Look-Ahead Processing mode (CVE-2026-56208), potentially causing denial of service or arbitrary code execution. Three additional flaws in spatial and temporal layer ID validation in the SVC encoder controls (CVE-2026-56209, CVE-2026-56210, CVE-2026-56211) allow arbitrary memory writes, out-of-bounds heap reads, or code execution. Users should apply the updated packages.

Ubuntu Security Notices·Sep 16, 09:29 UTC · 21h agoAdvisoryCVE-2026-56208CVE-2026-56209CVE-2026-56210+1 CVEs

Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs

Fortunate Recall introduces ontology-based lifecycle policies for LLM memory, cutting confabulation roughly in half (e.g., 45.1% to 22.4%) versus Mem0.

Fortunate Recall (FR) is a composable policy layer that classifies personal facts into a 10+1 behavioral ontology and applies category-specific lifecycle rules including differential temporal decay, slot-key supersession, event-time validity, and retrieval routing. FR-Bank scores 76.9% on the new 516-question LifecycleBench, ahead of Mem0, A-MEM, Memory-R1, and MemoryOS (61%-70.5%), and 75.2% on LongMemEval-S. End-to-end, confabulation drops from Mem0's 45.1% to 22.4% over answered queries, with the ranking replicating on open-weight Kimi K2.5 and transferring to the independent BEAM benchmark (46.8% vs 32.9%).

arXiv cs.AI / cs.LG / cs.CL·Sep 9, 16:31 UTC · 7d agoAI research

Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents

Researchers propose provenance-guided selective replay letting LLM agents forget revoked information without restarts, matching full reset behavior.

The paper formalizes execution-state unlearning for stateful LLM agents, requiring that agents behave as if a revoked memory record was never observed across transcripts, compressed memory, tool plans, and KV caches. It proves exact unlearning requires at least T-τ+1 recomputed transitions and that Provenance-Guided Selective Replay attains this bound via a provenance graph, KV cache cropping, and sanitized replay. In audits across three agent suites, nine baselines, and three model families, memory deletion left leakage unchanged, instruction-based forgetting collapsed under elicitation (Leak@probes = 1.00), and selective replay matched full resets at up to 9x fewer recomputed tokens.

arXiv cs.CR·Sep 4, 08:31 UTC · 12d agoAI safety & security1

"Shut Up and Let Me Enjoy My Otome": Understanding and Measuring the Toxicity in Otome Game Communities

First large-scale study finds 22.20% of Weibo otome game posts toxic versus 3.71% on Reddit, with LLM detectors reaching 0.82 F1.

Researchers present the first large-scale measurement of toxicity in otome game communities, introducing OtomeSCAN, which collected and analyzed 620,045 posts from Weibo and Reddit over 18 months. They manually annotated 4,308 posts, identified eight target groups, and evaluated seven toxicity detectors, with their best LLM-based model reaching F1-scores of 0.82 on Weibo and 0.78 on Reddit. The study found 22.20% of Weibo posts were toxic versus 3.71% on Reddit, and toxicity rose to 37.09% within 72 hours during an external attack on Weibo. The authors also flagged 191 potential-coordination clusters, 64.40% of which targeted game developers.

arXiv cs.CR·Sep 7, 21:44 UTC · 9d agoResearch

When the prompt becomes the payload: A practical pen-testing guide for GenAI, LLM and RAG applications

CSO Online publishes a practical penetration-testing guide for GenAI, LLM, and RAG applications, covering prompt injection, retrieval poisoning, and tenant isolation testing.

The guide frames LLM applications as attack graphs spanning prompts, retrieval layers, vector stores, tools, identities, and downstream APIs, arguing that conventional web testing misses instruction-vs-data channel risks. It builds on OWASP prompt injection guidance (direct vs. indirect injection) and NIST's 2025 adversarial machine-learning taxonomy, noting that RAG and fine-tuning do not remove injection risk. Recommended practices include documenting trust transitions across components, using canaries and synthetic records to avoid test side effects, running multi-turn and obfuscated injection campaigns, and verifying chains from poisoned documents to observable state changes. It also details testing RAG pipelines via controlled document poisoning across metadata, OCR layers, and code comments, plus cross-tenant isolation checks on retrieved document IDs.

CSO Online·Sep 9, 10:00 UTC · 7d agoAI safety & security1

Corrupt Plans, Clean Traces: Evading Chain-of-Thought Monitoring with Plan Injection

Plan injection plants benign-sounding harmful reasoning that steers LLM actors to adversarial actions while evading chain-of-thought monitors.

Researchers show that injecting harmful but benign-sounding plans into an actor model's context causes it to perform adversarial actions while its reasoning passes chain-of-thought monitors, achieving 25-33% monitor evasion rates across benchmarks and scaling to larger models like DeepSeek-R1. Actor models paraphrase injected plans as their own reasoning without attribution. Giving the monitor access to the injected plan dropped detection by up to 50% on the Bio-Math task, with extra thinking tokens spent rationalizing rather than flagging the plan.

arXiv cs.AI / cs.LG / cs.CL·Sep 14, 18:33 UTC · 2d agoAI safety & security

AmnesiaStealer Hijacks Chromium Sessions to Give Attackers Live Browser Control on macOS

Jamf details AmnesiaStealer, a new Rust-based macOS infostealer that hijacks Chromium sessions for live attacker-controlled browser access and steals Keychain data.

Jamf Threat Labs disclosed a three-stage macOS infostealer called AmnesiaStealer distributed via a fake GitHub download page using a ClickFix-style lure that tricks users into pasting a Base64 command into Terminal. The Rust payload harvests the Keychain, Apple Notes, Telegram sessions, Safari cookies, and 16 Chromium-family browsers, and captures the system password via a native prompt validated with dscl. A remotely fetched stream_module gives operators hidden, interactive browser control via the Chrome DevTools Protocol, streaming a live session at roughly 3fps. It persists via a root LaunchDaemon impersonating Apple's crash reporter and leverages the TCC bypass CVE-2020-9771 against macOS Catalina.

The Hacker News·Aug 15, 00:00 UTC · Aug 15, 2026Malware in the wildCVE-2020-9771

BdThemes Supply Chain Attack Poisons JSON to Create Rogue WordPress Admins

The Hacker News·Aug 11, 05:48 UTC · Aug 11, 2026VulnerabilityCVE-2026-18072CVE-2026-64638

A Malicious Webpage Could Poison Your Local AI Model Behind NVIDIA NemoClaw

Oasis Security found NVIDIA NemoClaw's Ollama binding to 0.0.0.0 enables DNS rebinding attacks that let attacker pages poison model chat templates with persistent hidden instructions.

Oasis Security disclosed that NVIDIA NemoClaw on Windows/WSL paths binds Ollama to 0.0.0.0:11434 without authentication, exposing the API to browser-based DNS rebinding attacks from malicious webpages. An attacker can then modify the model's chat template via /api/create, planting hidden instructions that run on every subsequent inference and persist across conversations, invisible to API consumers. NemoClaw v0.0.35 fixed the issue on macOS and Linux; no fix exists for Windows and WSL paths beyond a warning in v0.0.34. Ollama's own 2024 fix (CVE-2024-28224) added Host header validation, but it is skipped when bound to non-loopback addresses. No exploitation has been reported as of August 25, 2026.

The Hacker News·Aug 25, 15:58 UTC · 22d agoAI safety & securityCVE-2024-282241

Cognitive Extensions for Dual-Process Language Agents: Memory and Self-Reflection in Interactive Environments

Researchers extend the SwiftSage dual-process agent with adaptive memory and self-reflection modules, improving scores in interactive environments.

The work adds an Adaptive Memory Module (AMM) for salience-gated episodic storage and trigger-driven retrieval, and a Self-Reflection Module (SRM) for bounded execution-time validation and corrective intervention, to the SwiftSage agent. Controlled ablations on ScienceWorld across four configurations show the full system achieves the best mean final score (64.62), success rate (43.17%), and successful-step efficiency (19.33 steps). SRM is the strongest standalone contributor, suggesting execution-time control is the dominant bottleneck while episodic memory helps once the runtime loop is stable.

arXiv cs.AI / cs.LG / cs.CL·Sep 16, 17:50 UTC · 12h agoAI research

Quantifying the Engagement Trap: Impact of Short-form Video Recommender Systems on Users with ADHD

A 302-participant study finds engagement-optimized short-form video recommenders cause disproportionate time blindness and distress for users with ADHD.

Researchers ran a stratified Prolific study with 302 participants comparing short-form video recommendation experiences with and without ADHD. Participants with ADHD reported significantly higher time blindness, post-usage regret, and emotional distress despite perceiving recommendations as similarly relevant. The paper proposes neurodiversity-aware, human-centered design interventions to mitigate these algorithmic harms.

arXiv cs.AI / cs.LG / cs.CL·Sep 7, 17:38 UTC · 9d agoAI research

Fake bank websites play dead to evade security scanners

Fortra uncovered Chameleon SEO Poisoning: cloaked typosquat bank sites rank on Google and Bing to steal credentials while evading scanners, with cases up 40% in Q2 2026.

Fortra's threat intelligence unit FIRE spent three months tracking Chameleon SEO Poisoning, a phishing method that ranks recently registered typosquat domains on second-level domains like .ph.com and .gr.com above legitimate bank sites on Google and Bing. The technique uses presentation control (cloaking by referrer): direct visits get a dead, offline-looking page, while search-referred clicks receive a convincing fake bank login page, letting poisoned results persist for days or weeks. Fortra recorded a 40% jump in cases during the second quarter of 2026 and recommends referrer-spoofed URL testing, faster SLD takedowns, and bookmarking bank logins.

Help Net Security·Aug 24, 00:00 UTC · 24d agoPhishing & fraud in the wild

Untracked Nightmares: The Threats Hiding Behind Commodity Infrastructure

Unit 42 exposes CL-CRI-1171, a pay-per-install network spreading malware like Insomnia RAT via YouTube channels and SEO poisoning for over two years.

Palo Alto Networks Unit 42 details CL-CRI-1171, a cybercrime cluster operating a pay-per-install (PPI) marketplace that has delivered multiple malware families for at least two years. The group used at least eleven YouTube gaming channels with hundreds of thousands of followers, plus SEO poisoning promoting trojanized software such as a Bluetooth driver and WinDirStat, infecting gamers and corporate endpoints including critical infrastructure and government entities. A single shared loader delivered payloads including Insomnia RAT, ARKTunnel, Docro Hijacker, GCleaner and Socks5Systemz between July 2025 and April 2026, with more than 10,000 distinct loader samples and over 200 rotating C2 domains identified. YouTube terminated the malicious channels after Unit 42 notified the platform.

Palo Alto Unit 42·Sep 9, 10:00 UTC · 7d agoMalware in the wild1

Hackers deploy Linux rootkit on F5 BIG-IP APM devices, hiding web shell in memory

Sophos analyzed a Linux rootkit, tracked as PoisonedRefresh, on hacked F5 BIG-IP APM devices hiding an in-memory web shell linked to exploited CVE-2025-53521.

Sophos and ESET analyzed an implant (named PoisonedRefresh by ESET) deployed on compromised F5 BIG-IP APM appliances, which serves enterprise, financial, government, and public sector deployments. The second-stage payload intercepts __libc_start_main, hooks Apache's APR module loader (apr_dso_load), RC4-encodes key strings, and rewrites PHP file reads in memory to embed a web shell inside legitimate scripts like apm_css.php3, full_wt.php3, and webtop_popup_css.php3, leaving on-disk files unchanged. A related installer component infects /usr/sbin/httpd, persists across BIG-IP upgrade images, modifies SELinux configuration, and deploys the payload. The malware also creates a Unix domain socket at /run/bigtlog.pipe that hands authenticated local connections to /bin/bash; F5 links the activity to CVE-2025-53521, an unauthenticated RCE in BIG-IP APM already exploited in the wild.

Help Net Security·Sep 9, 11:05 UTC · 7d agoMalware in the wildCVE-2025-535211

New hardware device can RAM into encrypted memory, expose your data

Researchers built a $200 DDR5 interposer that silently drops memory writes to break TDX, SGX, and SEV-SNP confidential VM integrity, requiring physical access.

Researchers from KU Leuven, ETH Zurich, Durham University, and Google demonstrated DDRop, a hardware interposer costing under $200 that corrupts DDR5 bus commands to silently drop writes to encrypted memory, enabling replay attacks on confidential VMs. Because scalable memory encryption lacks freshness checks, protected VMs keep computing on stale attacker-selected data; on an Intel TDX server the attack forces debug mode for plaintext memory reads or forges attestation reports, succeeding in under two minutes without crashing. Intel and AMD both called the attack out of scope for their cloud threat models, with no mitigation planned, and proposed cache line versioning appears still vulnerable. The full interposer design is being released as open-source hardware.

The Register · Security·Sep 14, 18:31 UTC · 2d agoResearch

F5 BIG-IP APM Malware Injects a PHP Web Shell Into Memory, Evading Disk Scans

Sophos found F5 BIG-IP APM malware that injects a PHP web shell into Apache's memory, evading disk scans, linked to exploited CVE-2025-53521.

Sophos's September 7 analysis describes malware tied to F5's c05d5254 activity that hooks apr_dso_load, modifies the libphp PHP module in memory, and injects a PHP web shell when Apache loads apm_css.php3, full_wt.php3 or webtop_popup_css.php3, leaving on-disk files clean. A separate installer infects /usr/sbin/httpd and umount, runs before Apache starts, disables SELinux (per ESET's related PoisonedRefresh analysis), and opens a local socket at /run/bigtlog.pipe for shell access. The activity is linked to CVE-2025-53521 in BIG-IP APM, rated 9.8 CVSS 3.1, patched in October 2025, added to CISA KEV on March 27, 2026.

The Hacker News·Sep 9, 07:36 UTC · 7d agoMalware in the wildCVE-2025-53521

Bag of Tricks or Bag of Myths? Reducing Modeling Complexity with Task Knowledge in Explainable Suicide Risk Assessment

Researchers audit 31 NLP techniques for clinician-annotated suicide risk prediction, finding only 5 of 31 comparisons yield reliable gains.

A study of 1,635 clinician-annotated social media posts ran roughly 300 controlled experiments across 7 methodological families, auditing techniques such as model scaling, synthetic data, ensembling, and threshold tuning under severe class imbalance. The proposed system reformulates risk factor prediction as entailment between posts and codebook definitions, using architecturally diverse ensembles with class-balanced training and deployment-consistent calibration. It scores 0.8203 for risk, 0.7953 for evidence, and 0.7045 macro-F1 for factors, ranking third among 53 teams.

arXiv cs.AI / cs.LG / cs.CL·Sep 7, 17:14 UTC · 9d agoAI research

CISA Warns of Fortinet Heap-based Buffer Overflow Flaw Exploited in Attacks

CISA added actively exploited Fortinet CVE-2025-25249, a critical heap-based buffer overflow in FortiOS, FortiSwitchManager, and FortiSASE, to its KEV catalog.

CVE-2025-25249 is a heap-based buffer overflow (CWE-122/CWE-787) allowing unauthorized code execution by sending specially crafted packets. CISA added it to the Known Exploited Vulnerabilities catalog on September 9, 2026, with a September 12 remediation deadline for federal agencies under BOD 26-04 and mandatory forensic triage of affected environments. Internet-facing Fortinet firewalls and SASE platforms are a likely foothold for credential theft, persistence, and lateral movement; ransomware use is currently listed as unknown.

Cyber Security News·updated Sep 10, 18:57 UTC · 6d ago·first Sep 10, 11:26 UTC · 6d agoExploit / PoC in the wild 6 sourcesCVE-2025-252492

Bias-Induced Crossover in Absolute Capacity of Dense Associative Memory

Analysis shows biased patterns cut dense associative memory capacity from N^(n-1)/ln N to O(N^(n/2)), with a bias-induced crossover.

The paper analyzes dense associative memory capacity for biased centered binary patterns under the Krotov-Hopfield single-site criterion. Unbiased patterns (q=1/2) with order-n polynomial interactions yield capacity of order N^(n-1)/ln N, while fixed bias q<1/2 reduces capacity to O(N^(n/2)) for even n>=4 and O(N^((n+1)/2)) for odd n>=5. A bias-dependent crosstalk mean destabilizes sites carrying the frequent value, and an activity-dependent control potential restores the higher capacity within the conditioned-Gaussian approximation.

arXiv cs.AI / cs.LG / cs.CL·Sep 15, 17:16 UTC · 1d agoAI research

Heap overflow in kernel driver due to missing size validation

Fortinet fixes a CVSS 7.3 heap overflow in the FortiClient Windows kernel driver enabling code execution via crafted DNS responses.

Fortinet PSIRT advisory FG-IR-26-156, revised 2026-08-12, describes a heap-based buffer overflow (CWE-120, buffer copy without checking input size) in the FortiClient Windows kernel driver, scored CVSSv3 7.3. An unauthenticated attacker positioned to alter or craft DNS responses for a targeted host could execute arbitrary code via malicious packets. No CVE identifier or exploitation status is provided in the advisory text, so administrators should check the full bulletin for affected versions and fixed releases.

Fortinet PSIRT·Aug 12, 07:00 UTC · Aug 12, 2026Advisory

SimpleMemVLA: A Simple but Effective Native-Video Memory for Vision-Language-Action Models

SimpleMemVLA passes full timestamped video history straight to a VLA backbone, setting state of the art on four memory benchmarks.

SimpleMemVLA is a vision-language-action model for long-horizon manipulation that removes the dedicated memory module entirely. It keeps sampled history intact and feeds it to the backbone as timestamped video, with the hidden states of a generated sub-task serving as the only channel into a standard flow-matching action head. Prefilling the shared history prefix during action execution keeps latency close to a single-frame VLA. The system sets a new state of the art on four memory benchmarks and outperforms retrieval, compression and recurrent-state mechanisms, with causal interventions confirming the policy genuinely reads its history.

Hugging Face daily papers·Sep 1, 20:00 UTC · 15d agoAI research

Continual Learning Mechanisms Compose for Long-Horizon Memorization

Composing data, function, and weight anchors with merged LoRA raises 100-task long-horizon retention from 1.2% to 34.9% in continual fine-tuning.

The paper introduces long-horizon memorization: a model learns 100 query-answer tasks through continual supervised fine-tuning without retaining earlier examples or receiving task identifiers at inference. No single continual learning mechanism maintains strong retention at this horizon, so the authors compose complementary mechanisms along data/function/weight anchors and low-rank allocation rules. The best method combining all three anchors with merged LoRA ranks among the top 3 methods on all three datasets and raises average final retention from 1.2% to 34.9%, a 28-fold improvement.

Hugging Face daily papers·Sep 6, 20:00 UTC · 10d agoAI research

Hiding Prompt Injection in Legal Filing

A judge banned a plaintiff from electronic court filings after hidden prompt-injection text was discovered planted in legal documents.

Bruce Schneier's blog discusses an incident in which hidden prompt-injection instructions were planted inside a legal filing, apparently targeting AI systems that might process court documents. Judge Walter Spader Jr. responded by banning the plaintiff from electronic filings, requiring all future submissions as printed hard copies. Commenters debate whether the tactic could affect future AI-based processing of court records and whether plain-text formats will regain favor.

Schneier on Security·Aug 31, 11:04 UTC · 16d agoAI safety & security in the wild