Search: “ClearFake”

ClearFake Deploys Crypto Stealer That Uses Vulnerable Driver to Kill EDR Security Tools

Cisco Talos details ClearFake's fake-CAPTCHA chain deploying ZigCryptoStealer with a BYOVD attack that kills EDR processes, observed at a Ukrainian government organization in April 2026.

ClearFake compromises websites, injects JavaScript via a malicious Cloudflare Worker, retrieves instructions from BNB Smart Chain contracts (EtherHiding), and presents a fake Google CAPTCHA that tricks Windows users into pasting a command that loads a remote library over WebDAV via rundll32. The crypto-stealer branch uses DLL side-loading with a signed Chrome component to launch ZigCryptoStealer, which hijacks clipboard cryptocurrency addresses, alongside a signed but vulnerable Windows driver used in a BYOVD attack to terminate EDR processes. A parallel branch delivers Amatera secondary payloads that install a hidden remote-access client providing operator desktop control, with Cisco Talos tracking the remote-loader activity as UAT-10820. Talos observed unusual remote library execution at a Ukrainian government organization in April 2026 and assesses the attacks are part of a broader theft operation rather than a single targeted campaign.

Cyber Security News·Sep 9, 14:26 UTC · 7d agoMalware in the wild1

ClearFake WebDAV infection chain delivers Amatera stealer, ZigCryptoStealer, and NetSupport Manager

Cisco Talos details ClearFake WebDAV chains delivering Amatera stealer to a Ukrainian government organization, with cryptocurrency and credential theft payloads.

Cisco Talos investigated DLL executions named 'verification.google' via WebDAV UNC paths at a Ukrainian government organization, tracking the actor as UAT-10820 and assessing with moderate confidence the activity is Russian and opportunistic rather than targeted. The infection chain uses ClearFake JavaScript injected via a Cloudflare Worker, EtherHiding storage on BNB Smart Chain contracts, and a ClickFix fake Google CAPTCHA prompt to deliver Amatera stealer. Secondary payloads differ by C2: one loader deploys ZigCryptoStealer with a Go reverse TCP proxy and a vulnerable driver that kills EDR, while the other installs an unauthorized NetSupport Manager with a Russia-based C2. Similar Amatera chains were separately documented by Malwarebytes and Blackpoint Cyber, but with no shared infrastructure.

Cisco Talos·Sep 8, 10:01 UTC · 8d agoMalware in the wild1

PuzzleMask: Abusing Plain Prose as a Covert AI Attack Vector

Check Point details PuzzleMask, a plain-prose technique that bypasses LLM gatekeeper policy checks, letting hidden payloads reach target models unreviewed.

Check Point Research describes PuzzleMask, a prompt-crafting technique that hides policy-violating payloads inside plain-English prose wrappers, bypassing quick LLM-based policy checks without emojis, Base64, or invisible formatting. The researchers tested 23 automated prompts against gatekeepers including GPT-4o-mini, GPT-OSS-Safeguard 20b, Claude 3 Haiku, and Llama Guard 3, and all were classified as safe despite policies that flagged the plain versions. When submitted to GPT-5 in thinking-high mode with a Python interpreter, the target model extracted and acted on the payload in over 90% of trials. The technique is not itself a jailbreak but can carry a jailbreak prompt as payload; mitigations include input paraphrasing, hardened gatekeeper policies, and output monitoring.

Check Point Research·updated Sep 11, 05:53 UTC · 5d ago·first Sep 10, 14:32 UTC · 6d agoAI safety & security 2 sources

The Top 4 Threats We Found by Investigating Every Alert for a Quarter

Prophet Security's quarterly threat report finds identity attacks made up roughly half of confirmed malicious activity, with stolen authenticated sessions bypassing conditional access controls.

Between May 1 and July 31, 2026, Prophet Security investigated every alert in customer environments immediately on arrival; about 7% of completed investigations were confirmed malicious. Direct account/session attacks made up roughly 18% of confirmed activity, with replayed authenticated sessions bypassing conditional access since no re-authentication occurs. Infostealer activity affected about a quarter of investigated organizations, mostly delivered via browsers through compromised websites, malicious ads, sponsored search results and ClickFix fake CAPTCHA gates, with payloads including Lumma Stealer, Vidar, HijackLoader and AsyncRAT. Credential phishing was the largest category at about 28%, with adversary-in-the-middle relays revealed by seconds-long MFA approvals from distant countries and MFA fatigue attacks exploiting automatic unlock policies.

BleepingComputer·Sep 10, 14:00 UTC · 6d agoThreat actor in the wild

Reverse-Lookup Service Exposed Millions of Photos of People’s Faces

People-search service ClarityCheck exposed a database with more than 9 million facial image files despite marketing its reverse lookup as private and secure.

WIRED reports that ClarityCheck, a people-search tool offering reverse image lookup, left a database containing more than 9 million image files of people's faces exposed. The company markets the service as private and secure. The exposure affects millions of individuals whose facial photos could be linked to their identities.

WIRED · Security·Aug 19, 10:00 UTC · 28d agoData breach

WordlistLoader Delivers Amatera via ClickFix, SynkLoader Phishes Windows Passwords

Gen Digital details WordlistLoader delivering Amatera Stealer via ClickFix and EtherHiding, while SynkLoader phishes Windows credentials through Microsoft Teams.

Gen Digital identified WordlistLoader, an intermediate stage that reconstructs shellcode encoded as plain English words (or 16-byte UUID chunks) and loads Amatera Stealer 4.3.3-alpha1, delivered through ClearFake ClickFix prompts on compromised websites. The chain uses EtherHiding to fetch JavaScript from a blockchain smart contract, hidden cmd.exe via conhost, WebDAV-mounted shares with rundll32 execution, and ETW bypass via hardware breakpoints. The stealer adds hardened WoW64 syscalls, Heaven's Gate indirect-syscall trampolines and a redesigned application-bound encryption bypass. Separately, Expel observed SynkLoader distributed via Microsoft Teams phishing impersonating IT service desks, installing an MSI from Azure blob storage and serving a fake lock screen to capture credentials.

The Hacker News·Aug 26, 05:08 UTC · 21d agoMalware

Safe word: What is it and why do you need one?

ESET recommends pre-agreed family safe words to counter AI voice-clone scams such as virtual kidnapping, as one-in-four Americans report receiving deepfake calls.

ESET outlines how scammers use just seconds of audio scraped from social media or work content to create convincing voice clones, with a Hiya report finding one-in-four Americans received a deepfake voice call in the past 12 months. Common schemes include virtual kidnapping calls mixing cloned voices with sobbing and background noise. A pre-agreed, non-OSINT-discoverable safe word, plus callback verification via known numbers and 2FA, reduces success rates of these frauds.

ESET WeLiveSecurity·Sep 9, 09:00 UTC · 7d agoPhishing & fraud in the wild

Risky Bulletin: Slovakia finds Russian backdoor in traffic speed cameras

Slovakia's NBU found an SMS-triggered backdoor in Russian-made NERO R-ONE traffic cameras, pausing a 279-unit deployment.

Slovakia's national security service NBU issued an alert against NERO R-ONE high-speed traffic cameras after finding a backdoor that grants shell and network access via SMS from hardcoded Russian phone numbers. The cameras are a rebranded version of the Russian CORDON PRO.M model by St. Petersburg firm Semicon, purchased via a Cyprus shell company under a €30 million EU-funded project. The report also found SecureBoot disabled, vulnerable web management, and unauthenticated live streams; the Interior Ministry paused deployment of 279 cameras pending independent assessment.

Risky Business News·Aug 19, 04:52 UTC · 28d agoThreat actor in the wild1

Adaptive Gated Deepfake Detection for Low-Resolution and Resource-Constrained Environments

AdaGate-DF routes deepfake detection by image quality through dual multi-exit gates, hitting 0.9370 AUC on Celeb-DF with low inference latency.

AdaGate-DF is an adaptive gated deepfake detection framework that uses image-quality cues to send high-quality images through earlier exits, saving compute in resource-constrained settings. On Celeb-DF it achieves an AUC of 0.9370, outperforming MaD-CoRN and DefakeHop++, and reaches 0.9708 at 384x384 resolution. On FaceForensics++ it remains effective under class imbalance while balancing uncertainty-aware prediction and computational efficiency.

arXiv cs.AI / cs.LG / cs.CL·Sep 4, 16:14 UTC · 12d agoAI research

Hackers Abuse Google CAPTCHA, WebDAV and BNB Smart Chain to Deploy Credential-Stealing Malware

Cisco Talos details ClickFix campaigns abusing fake Google CAPTCHA prompts, WebDAV and BNB Smart Chain to deploy the Amatera infostealer.

Cisco Talos identified a multi-stage campaign, first observed in April 2026 when a Ukrainian government organization executed a disguised DLL named verification.google from a WebDAV path. The chain combines ClearFake JavaScript injected via malicious Cloudflare Workers, EtherHiding payloads stored in BNB Smart Chain contracts, ClickFix social engineering, and WebDAV-hosted DLL loaders to deliver the Amatera (ACR Stealer) infostealer. The Amatera configuration contained over 400 collection entries targeting browsers, messengers, crypto wallets, password managers, FTP and VPN tools, hunting for .kdbx, .p12, .pfx and .pem files. The pf.ch branch deploys ZigCryptoStealer, an EDR-terminating vulnerable driver and Go-based reverse TCP proxies, while the verification.google branch installs NetSupport Manager, supporting moderate-confidence attribution of the UAT-10820 activity to a Russian actor.

GBHackers·Sep 9, 13:27 UTC · 7d agoMalware in the wild

I’ve been deepfaked: What do I do?

ESET outlines steps for deepfake victims: preserving evidence, using platform reporting tools, and legal remedies like the US TAKE IT DOWN Act and StopNCII.org.

ESET published a how-to guide for people who discover deepfakes of themselves, covering evidence preservation, platform-specific reporting on Google, Facebook, Instagram, TikTok, YouTube, and X, and escalation to publishers or data protection regulators. It notes the US TAKE IT DOWN Act criminalizes non-consensual intimate imagery (NCII) and requires 48-hour takedowns, while UK and EU laws add creation offenses and GDPR Article 17 erasure rights. Services like StopNCII.org and TakeItDown.NCMEC.org hash images so participating platforms such as Meta, TikTok, Reddit, and X can find and remove matching copies.

ESET WeLiveSecurity·Sep 2, 11:48 UTC · 14d agoAI safety & security1

E4del and PINHOLE RATs Turn FTP Banners Into Dead Drops for Malware Commands

Researchers identified two new RATs, E4del and PINHOLE, using FTP banners as dead drop resolvers to fetch C2 commands via ClickFix-style lures.

SOCRADAR documented two previously unreported RATs, E4del and PINHOLE, that use FTP banners as dead drop resolvers, a first in-the-wild use of this technique. E4del is a Node.js-based RAT hidden in a digitally signed Electron application masquerading as Discord, with tiered-jitter beaconing, reverse shell, screenshot capture and desktop streaming. PINHOLE is more advanced, using Pinterest, SurveyMonkey and Cloudflare Workers for C2 resolution, six layers of unpacking, Halo's Gate EDR evasion and Early Bird APC injection. A related WebDAV delivery chain overlaps with ClearFake campaigns distributing WordlistLoader and Amatera Stealer.

The Hacker News·Aug 25, 11:33 UTC · 22d agoMalware in the wild

Sexually Explicit Deepfake Sites Target 100-Plus Politicians in Europe

Deepfake pornography sites have targeted nearly 150 European politicians, overwhelmingly women MPs, per new research on 160 abusive domains.

Researcher Benjamin Shultz analyzed roughly 160 deepfake abuse domains and found at least 138 women MPs from 22 EU countries appeared or were mentioned, versus nine male MPs — making women MPs 33 times more likely to be targeted. Sites host database-like profiles with names, photos, personal details, and links to 'nudifier' creation tools. The findings, published by German think tank Agora Digitale Transformation, show politicians from Germany, the Netherlands, Italy, and France most affected, with senior politicians targeted more often. The UK and EU are planning bans on nudify services, while the US Take It Down Act has taken major deepfake sites offline.

WIRED · Security·Sep 14, 11:00 UTC · 2d agoAI safety & security1

Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization

ModerationBench shows foundation models can nearly triple Bluesky's moderation F1 (0.60 vs 0.22), with instruction- and example-driven guidance performing comparably.

Researchers built ModerationBench, a new benchmark of 4,000 manually annotated in-the-wild posts from Bluesky, to test whether foundation models can reliably operationalize content moderation policies. They systematically compare instruction-driven guidance (reasoning from policy precepts) with example-driven guidance (generalizing from precedents) for Vision-Language Models. Both paradigms achieve comparable peak effectiveness, and foundation models nearly triple the F1 of Bluesky's deployed moderation system on Random Posts (0.60 vs 0.22).

arXiv cs.AI / cs.LG / cs.CL·Sep 9, 16:30 UTC · 7d agoAI research1

Hackers Can Hide Malicious AI Commands Inside Normal English to Bypass Security Filters

Check Point's PuzzleMask technique hides malicious prompts in ordinary English that fast gatekeeper models miss but high-reasoning downstream models execute.

Check Point researchers disclosed PuzzleMask, a technique concealing policy-breaking instructions in natural-language prose without encodings or invisible characters. Fast screening models classified all 23 crafted wrappers as safe, while a high-reasoning model recovered and acted on the hidden instruction in 17 of 18 tests (94.4%). The gap stems from capability imbalance between gatekeeper and target models, with defenses including paraphrasing untrusted input, stricter self-referential wording rules, and output/tool-call monitoring.

Cyber Security News·Sep 11, 07:36 UTC · 5d agoAI safety & security1

DeepSeek v4.1 Flash Uncensored

Hugging Face user dealignai published an uncensored FP8-quantized variant of DeepSeek v4.1 Flash, drawing moderate Hacker News attention.

A community-published uncensored FP8 quantization of DeepSeek v4.1 Flash appeared on Hugging Face under user dealignai. The release is a third-party upload rather than an official DeepSeek launch, and no benchmark data or license details are provided in the listing. It received limited visibility, with 43 points and 11 comments on Hacker News.

Hacker News · AI·updated Sep 16, 12:19 UTC · 13h ago·first Sep 11, 06:46 UTC · 5d agoModel release 3 sourcesHN 43↑ · 11 comments1

Deep-Fake CAPTCHA: Mitigating Next-Generation Social Engineering Attacks

Researchers propose DF-CAPTCHA, a challenge-response defense that verifies callers in voice and video to defeat real-time deepfake impersonation in social engineering.

The DF-CAPTCHA framework prompts call participants with simple challenge-response tasks that are easy for humans but hard for real-time deepfake systems to convincingly generate. Responses are verified on four criteria: realism, identity consistency, task completion, and response time. User studies and experiments with real-time deepfake models across audio and video modalities show substantially improved detection over passive artifact-based methods.

arXiv cs.CR·Sep 10, 11:35 UTC · 6d agoResearch1

AuK Technical Report: An Open-Source Foundational Model for Speech Generation and Editing

Open-source speech foundation model AuK unifies generation and editing, trained on 1.95 million hours, with distilled AuK-Flash achieving 4.5x speedup.

AuK is an open-source foundational model that unifies speech generation and editing through natural-language instructions and audio context, trained on approximately 3.03 billion instruction-audio instances and 1.95 million hours of supervision across five task families including generation, content editing, and acoustic editing. It combines a multimodal LLM for semantic conditioning, a VAE jointly trained on speech, general audio, and music, and a hybrid rectified-flow Transformer using dual-stream MMDiT blocks followed by unified single-stream DiT blocks. Post-training applies human-feedback preference optimization for editing and reward-based reinforcement learning for generation, and the distilled AuK-Flash performs 4-step inference without classifier-free guidance at a 4.5x wall-clock speedup. Source code and model weights are released.

Hugging Face daily papers·Sep 7, 20:00 UTC · 9d agoModel release2

When Models Edit Too Much: On the Fidelity of Minimal Code Edits

A 400-task BigCodeBench evaluation shows frontier LLMs widely over-edit code; a preservation instruction cuts excess edits and raises Pass@1 by 2.3 points.

Researchers built an evaluation framework from 400 BigCodeBench problems with injected AST-level corruptions, each with a known minimal patch, to measure over-editing in LLM code repair. Even strong models like GPT-5.5 produce unnecessarily large edits despite high Pass@1. Adding a preservation instruction reduced average excess Levenshtein distance from 0.195 to 0.131, cut added cognitive complexity by 26.6%, and raised Pass@1 by 2.3 points. Reinforcement learning post-training gave the best out-of-domain edit-fidelity trade-off, while supervised fine-tuning overfit to seen corruption patterns.

Hugging Face daily papers·Sep 2, 20:00 UTC · 14d agoAI research1

DF26: We Cannot Tell Fake From Real Anymore

DF26 benchmark shows humans and state-of-the-art deepfake detectors perform near chance on videos generated by seven modern text-to-video models.

Researchers introduce DF26, a benchmark of 271 real and 2,420 fully synthetic videos created by seven modern video generation models, all depicting single-person public-speaking scenarios such as direct-to-camera recordings, official statements, and studio interviews. Human viewers and state-of-the-art deepfake detectors scored close to random chance at distinguishing fakes from real footage. The authors argue current evaluation protocols are insufficient and call for benchmarks that explicitly measure robustness to modern generative model distribution shifts.

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

TCRF taken offline by DDoS attack after Claude user ban

The Cutting Room Floor game wiki was taken offline by a DDoS attack after a user leveraging Claude was banned.

The Cutting Room Floor (TCRF), a wiki documenting unused video game content, was knocked offline by a distributed denial-of-service attack. The attack reportedly followed moderation action banning a user who was using Anthropic's Claude. The incident highlights friction between community sites and AI-assisted users and tools.

Lobsters · security·Aug 28, 21:26 UTC · 19d agoAI safety & security in the wild

Omni-Streaming Thinking

Omni-Streaming Thinking fixes premature cross-modal commitment in streaming omni-modal models via pending claims verified against modality-specific evidence, beating baselines by over 10%.

The paper identifies 'premature cross-modal commitment', where streaming models keep relaying early visual interpretations even after audio contradicts them. OST generates evidence-linked pending claims with future verification intervals, stores audio and visual evidence separately, and refutes claims when contradictory evidence appears. Built on a frozen Qwen3-Omni-30B-A3B-Instruct backbone with lightweight adaptation, it outperforms open baselines by more than 10% relative on five streaming and audio-visual benchmarks. On the new OST-DiagBench it reaches d-prime 2.95 versus at most 1.38 for open baselines, while reducing vision-induced auditory hallucinations.

Hugging Face daily papers·Sep 13, 20:00 UTC · 3d agoAI research1

One Editor, Many Edits: A Unified Training-Free Framework for Diverse Video Editing

Researchers introduce EditVid, a training-free video editing framework scoring 78.16 FiVE-Acc versus 58.95 for the strongest comparable baseline.

EditVid is a unified training-free framework for diverse instruction-guided and subject-guided video editing. It combines sparse causal memory for local coherence, correspondence-based post-attention token injection for long-range identity preservation, and soft latent blending for edit locality. On the FiVE benchmark it reaches 78.16 FiVE-Acc against 58.95 for the strongest evaluated training-free baseline, with competitive results on IVEBench. A user study showed 51.8% overall preference for EditVid over 7 competing methods.

Hugging Face daily papers·Sep 2, 20:00 UTC · 14d agoAI research

LexFlip: A Dissociation Diagnostic for Legal Meaning Preservation Metrics

LexFlip releases 373 minimal perturbations of Quebec statutory French that reverse legal force while preserving tokens, exposing weaknesses in embedding-based meaning preservation metrics.

LexFlip provides 373 minimal perturbations of Quebec statutory French that reverse legal force while preserving 0.93 of tokens, creating dissociation items that break monotone token-overlap metric validation. The seven embedding and BERTScore metrics tested register only 0.022-0.039 of their identical-to-unrelated range on these edits, versus 0.670 for bidirectional NLI. Against FrJudge, with a measured human ceiling of r=0.597, a bare length feature outscores every semantic metric tested.

arXiv cs.AI / cs.LG / cs.CL·Sep 4, 15:49 UTC · 12d agoAI research

ClickFix Campaigns Abuse Legitimate Services for Persistent Access

Dark Reading reports two ClickFix social engineering campaigns abusing legitimate services to compromise organizations and maintain persistent access.

Dark Reading describes two separate attacks in which threat actors used the ClickFix social engineering tactic to compromise organizations. The campaigns abuse legitimate, trusted services to gain and maintain persistent access to victim environments. No specific victims, actors, or indicators were named in the available text.

Dark Reading·Sep 8, 17:25 UTC · 8d agoThreat actor in the wild

New York Seizes a Dozen Celebrity Deepfake Websites

Manhattan DA seized 12 domains hosting nonconsensual celebrity deepfake videos depicting about 1,200 victims, the largest such takedown to date.

The Manhattan District Attorney's Office seized 12 domains used for sharing and selling nonconsensual deepfake sexual videos, depicting around 1,200 people, mostly women, including influencers, politicians, and celebrities. The seizures under New York criminal procedure law represent the largest takedown of explicit deepfake websites to date, aided by the US Take It Down Act. Much of the content appeared to be reposts from the MrDeepFakes platform, and investigations into operators and uploaders are ongoing.

WIRED · Security·updated Sep 14, 17:26 UTC · 2d ago·first Sep 14, 16:50 UTC · 2d agoPolicy & legal 2 sources

Motion-Omni: End-to-End Joint Speech and Full-Body Motion for Spoken Dialogue

Researchers introduce Motion-Omni, an end-to-end model generating speech with synchronized full-body motion, responding 5.4x faster than cascade pipelines.

Motion-Omni is an end-to-end framework in which a spoken dialogue model outputs facial expressions and hand, upper-body, and lower-body motion directly from the hidden states that produce speech, replacing two-stage cascade pipelines. Trained on 422,856 quality-ranked pseudo-labeled pairs (1,402 hours) with a Qwen2.5-7B-Instruct backbone, Motion-Omni-Q7 matches its teacher cascade within 2% on reference-free motion metrics, achieves a 2.62% word error rate, and runs faster than real time (RTF=0.78). The authors also release the SwDA-500 dataset and the first public evaluation protocol for stochastic open-ended full-body spoken dialogue.

Hugging Face daily papers·Aug 27, 20:00 UTC · 20d agoAI research1

Rare Not Random Using Token Efficiency for Secrets Scanning

Researcher proposes token efficiency (string length divided by BPE token count) as a better post-regex filter than entropy for secrets scanning, validated on CredData.

The post explores whether Byte-Pair Encoding tokenization can replace Shannon entropy as the primary filter for candidate secrets captured by regex in tools like Gitleaks. It defines 'token efficiency' as string length divided by token count under the cl100k_base tokenizer; secret-like strings such as GitHub tokens tokenize into many small tokens and score low, while natural text scores high. Evaluating labeled secrets from the CredData dataset shows a usable separation, with roughly 2.5 suggested as a minimum cutoff versus Gitleaks' 3.5 entropy threshold. The technique is positioned as a post-regex filtering step rather than a standalone detector.

Lobsters · security·Sep 12, 06:42 UTC · 4d agoResearch

Enoki: Efficient Multi-Level Hallucination Detection

Researchers introduce Enoki, an open information extraction framework unifying claim-level and span-level hallucination detection in LLMs at lower inference cost.

Enoki extracts text-anchored relational facts, verifies them against evidence, and projects unsupported facts back onto hallucinated spans, so claim-level verification and span-level localization share one representation without separate alignment. It supports LLM-based, encoder-based, and rule-based extraction regimes to balance accuracy against inference cost. Experiments show it stays competitive with strong claim-level systems while using fewer resources and outperforms them on fine-grained span- and entity-level localization. The authors also release EnokiQA, a dual-granularity dataset with aligned claim-level verification and span-level localization annotations.

Hugging Face daily papers·Aug 31, 22:21 UTC · 16d agoAI research

Novel-reading apps used users' phones to generate fake ad traffic

Help Net Security·Aug 6, 00:00 UTC · Aug 6, 2026Threat actor1

ENCP: Episode-Normalized Conformal Prediction for Vision-and-Language Navigation

ENCP calibrates conformal prediction per navigation episode, giving step-level coverage guarantees for vision-language navigation agents despite within-episode dependence.

Episode-Normalized Conformal Prediction (ENCP) rescales a nonconformity score by a VLN policy's residual confidence and calibrates one maximum score per episode, preserving step-level coverage of at least 1−α despite dependence among steps within an episode. Across four VLN policies and three nonconformity scores on R2R and REVERIE, ENCP meets all reported empirical step-coverage targets in seen-to-unseen evaluation. The model-agnostic uncertainty estimates can signal when an agent should defer to a stronger predictor or human assistance.

arXiv cs.AI / cs.LG / cs.CL·Sep 15, 17:42 UTC · 1d agoAI research

deepseek-ai/DeepSeek-V4.1-Flash — new model trending #28 on Hugging Face

DeepSeek releases DeepSeek-V4.1-Flash, a 552B-parameter multimodal MoE model with 1M-token context and KV cache cut to 890 bytes per token.

DeepSeek-V4.1-Flash is a multimodal Mixture-of-Experts model with a 552B-parameter backbone that activates 8B parameters per token during prefill and 16B during decode. It uses a Causal Encoder-Decoder architecture, Compressed Sparse Attention 2, and FP4 KV caching to reduce the global KV cache footprint to 890 bytes per token, roughly one quarter of DeepSeek-V4-Flash. The model was trained from scratch on 45T tokens with context extended to 1M tokens, includes an Engram conditional-memory module (196B parameters), and is released under the MIT license. Post-training uses SFT, RL, and on-policy distillation with large-scale automated synthesis of agentic tasks and a controllable reasoning effort setting from 1 to 100.

Hugging Face trending models·updated Sep 12, 05:56 UTC · 4d ago·first Sep 10, 02:17 UTC · 6d agoModel release 4 sources1

Knowing What Not to Answer: Selective Non-Compliance in Vision-Language Models

Researchers introduce KoNA, a benchmark exposing vision-language models' failures at selective non-compliance, plus fine-tuning that improves refusal and abstention accuracy.

KoNA is a benchmark for evaluating selective non-compliance in vision-language models across five categories: False Premise, Visual Inaccessibility, Universal Unknown, Task Feasibility and Safety. It tests both query-level and component-level non-compliance using paired single and compound queries, and evaluations across diverse VLMs show models often fail to refuse, correct or abstain appropriately, with failures worsening on compound queries. Fine-tuning VLMs on KoNA examples substantially improves non-compliance accuracy while largely maintaining performance on fully answerable tasks.

Hugging Face daily papers·Sep 3, 20:00 UTC · 13d agoAI research1