Search: “LLM CLI”

llm 0.34

Version 0.34 of Simon Willison's llm CLI adds response-duration metrics to log output, plus bug fixes and faster log querying.

The open-source llm command-line tool for interacting with large language models released version 0.34. The headline change adds response duration in milliseconds and human-readable form to llm logs --usage Markdown output, plus a new duration_ms field in llm logs --short. The release includes several contributed bug fixes and a significant performance improvement to llm logs, alongside the related llm-openrouter 0.7.1 release.

Simon Willison·Sep 2, 19:23 UTC · 13d agoAI tools & infra

llm-openrouter 0.7.1

Simon Willison released llm-openrouter 0.7.1, a performance fix for loading OpenRouter models in his LLM CLI plugin.

Version 0.7.1 of the llm-openrouter plugin addresses a performance problem when loading OpenRouter models in the LLM command-line tool. The fix was contributed by GitHub user waveplate. It is a minor maintenance release with no security implications noted.

Simon Willison·Sep 2, 20:23 UTC · 13d agoAI tools & infra

llm 0.35

llm CLI tool version 0.35 adds support for OpenAI's new GPT-6 Astra model exposed as gpt-6-astra.

Simon Willison released llm 0.35, which adds an OpenAI model definition for GPT-6 Astra under the model ID gpt-6-astra. No other release details were provided in the post.

Simon Willison·Sep 7, 23:54 UTC · 8d agoAI tools & infra

Re: AI slops from Eve

Solar Designer finds little similarity between recent LLM-generated fake advisories on oss-security beyond missing Date headers.

A mailing-list thread discusses 'AI slops', LLM-generated fake security reports posted to oss-security. Solar Designer notes the suspicious messages share only trivial traits like a missing Date header and LLM use. He concludes there is no significant problem with any particular sender or model and no investigation is needed.

oss-security·Sep 13, 15:08 UTC · 2d agoIndustry 12 sources1· 1 read

When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control

Chain-of-Self-Questioning prompting cuts LLM wrong-answer commitments 32% relative while raising answered accuracy, holding across eleven model families.

The paper introduces Chain-of-Self-Questioning (CoSQ), a prompt-only framework that makes LLM answer commitment conditional on an explicit assessment of the information required to answer. On an 817-item TruthfulQA multiple-choice set, Grounded-CoSQ at τ=0.90 reduced mean unconditional wrong-commitment rate from 13.1% under chain-of-thought to 8.9% (a 32.1% relative reduction), while raising answered accuracy from 86.9% to 89.7% at 87.6% coverage. Improvements held across eleven open-weight and hosted model families and at every evaluated threshold, with convergent evidence from a Natural Questions short-answer evaluation.

arXiv cs.AI / cs.LG / cs.CL·Sep 15, 17:52 UTC · 15h agoAI research

Amazon Kiro Prompt Injection Can Exfiltrate Sensitive Data Through Kiro Powers

Mindgard found a prompt injection flaw in Amazon Kiro IDE letting attacker-controlled workspace files exfiltrate sensitive local data; fixed in version 0.8.140.

Mindgard disclosed a prompt injection flaw in Amazon Kiro, an agentic AI IDE, that lets attacker-controlled repository content steer the agent into exfiltrating sensitive workspace data through Kiro Powers, which bundles MCP server configurations, POWER.md steering files, hooks, and contextual knowledge. Exploitation requires the user to open a malicious project via a workspace file and send any message to the agent; difficulty is rated low and it works in both trusted and untrusted workspaces. Amazon fixed the issue in Kiro IDE 0.8.140; the flaw has no CVE identifier and follows earlier Kiro bugs including CVE-2026-10591, plus related prompt-injection and code-execution issues in Codex CLI, Cursor, Gemini CLI, Copilot CLI, and Claude Code.

The Hacker News·Aug 28, 18:35 UTC · 18d agoAI safety & securityCVE-2026-10591CVE-2026-35603CVE-2026-257251

Can LLMs Engineer Their Own Agent Harness? ByteDance Seed’s HarnessDev Says Only 34 of 64 Changes Generalize

ByteDance Seed's HarnessDev benchmark finds LLM-built agent harnesses trail human engineering on code and search, with only 34 of 64 revisions generalizing.

Researchers from ByteDance Seed, SUTD, Georgia Tech, M-A-P, and TokenWave.AI introduce HarnessDev, a benchmark that evaluates the runnable agent harness an LLM writes rather than its answers, using Creation and Evolution stages across SWE-bench Pro, Terminal-Bench 2.1, MLE-bench, EQ-Bench3, and BrowseComp (2,207 instances). Six creator models including Opus 4.8, GPT-5.5, Gemini 3.1 Pro, DeepSeek V4 Pro, Qwen 3.7 Max, and Seed 2.0 Pro were tested; Opus 4.8 posted the best average of 67.8 versus an 86.2 human-engineered reference. Self-built harnesses beat references on writing and ML experimentation but lag badly on code and search, and quality proved executor-specific: Opus 4.8 fell from 69.3 to 33.0 on SWE-bench Pro when the executor was switched to Gemini. Evolution gains were small and noisy: of 64 adjacent changes, feedback and held-out scores agreed only 34 times (53.1%), and much generated state and memory code never executed.

MarkTechPost·Sep 11, 22:01 UTC · 4d agoAI research 2 sources

Hackers Use LLMs to Generate Exploit Scripts and Automate Post-Exploitation Across Latin America

Unit 42 says Latin American attackers used LLMs to automate post-exploitation in campaigns hitting Mexican government, water utilities, and Brazilian financial firms.

Unit 42 identified two campaigns in Latin America whose operators used commercial LLMs (Claude, GPT-4.1) behind a self-hosted NextChat interface to generate and debug post-exploitation scripts. Cluster CL-CRI-1131 compromised a transportation organization, Mexican federal ministries, and water utilities in Mexico and Ecuador, using native Windows tools and Volume Shadow Copies to dump the SAM registry hive and NTDS.dit. Cluster CL-CRI-1163 targeted Brazilian financial organizations with job-themed phishing, custom RATs, and a Go-based reverse SOCKS5 tunneling utility called SockTz, with nine versions deployed within roughly two hours. Trend Micro tracks related AI-augmented activity as SHADOW-AETHER-040 and SHADOW-AETHER-064.

GBHackers·updated Sep 10, 08:06 UTC · 6d ago·first Sep 10, 06:02 UTC · 6d agoThreat actor in the wild 2 sources1

Autonomous AI Agents Compromise Thousands of Credentials in Under Six Hours

Google's GTIG reports threat actors using autonomous AI agents, credential stealers, and LLMs to steal AI models, API credentials, and harvest thousands of credentials.

Google Threat Intelligence Group says attackers are targeting proprietary AI models across healthcare, government, and media, exfiltrating API credentials, and co-opting victim cloud environments to run unauthorized AI workloads. TeamPCP (Altered Spider/UNC6780) is conducting supply chain compromises of PyPI, npm, and Docker Hub, deploying the SANDCLOCK and DUSTMAKER credential stealers, with DUSTMAKER using AI workspace poisoning and prompt injection for defense evasion. One financially motivated actor used an autonomous multi-agent framework to compromise thousands of third-party credentials in under six hours without human intervention. China-nexus groups UNC6508 and Basin Castle (Mustang Panda) used local open-weight LLMs and commercial LLMs like Gemini, Claude, and Codex for espionage tasks and evading provider monitoring.

The Hacker News·Sep 8, 13:48 UTC · 7d agoThreat actor in the wild 2 sources1

Verifiable by Construction: Claim-Level Evaluation of Verbatim Citation in Clinical Question Answering

Evaluation of twelve LLMs on 222 clinical questions shows verbatim quotes rarely substantiate claims; claude-opus-5 fully substantiates only 37.1%.

The authors build a standardized harness over four clinical practice guidelines and evaluate twelve LLMs on 222 synthetic clinical questions, measuring citation attachment, verbatim quote production, and claim substantiation. Most models attach verbatim quotes to over 90% of claims from prompting alone, though lightweight models like claude-haiku-4.5 struggle. Quotes frequently fail to substantiate claims: claude-opus-5 quotes 98.0% of claims but fully substantiates only 37.1%, exposing a capability gap for verifiable clinical QA.

arXiv cs.AI / cs.LG / cs.CL·Sep 14, 17:53 UTC · 1d agoAI research

OpenClaw Power, MacBook Simplicity: Five Days With Grok Bot

Hands-on review finds Grok Bot simplifies agent setup via browser logins and bot abstraction, contrasting with the user-owned OpenClaw platform.

After five days with Grok Bot, the reviewer highlights browser-based sign-in as the key differentiator: connecting X, Freshdesk, and Google Calendar required only logins, no MCP configs or API keys. The piece contrasts Grok Bot's managed 'agent computer' with OpenClaw 2.0's user-owned Gateway, which now supports reusing Claude Code or Codex logins and ships a native Codex runtime. Grok Bot introduces 'Bots' as composable units arranged in 'group chats', exemplified by an Agentic Engineer Bot routing tasks across Claude Code, Codex, and Grok Build CLI. The reviewer used Grok Bot with a Cursor Pro+ account.

Latent Space·Sep 5, 15:01 UTC · 10d agoAI industry1

llm-gemini 0.34

llm-gemini 0.34 adds support for Google's new Gemini 3.8 Flash model with configurable low, medium and high thinking levels.

Simon Willison released llm-gemini 0.34, a plugin for the LLM CLI that adds the gemini-3.8-flash model, including low, medium and high thinking levels, and fixes an issue where async responses failed to record the resolved model version. The release coincided with Google's launch of Gemini 3.8 Flash, plus a Gemini 3.8 Flash Cyber variant restricted to trusted defenders. Willison noted the Flash tier's speed, low cost and competence at HTML and JavaScript generation tasks.

Simon Willison·Sep 2, 16:39 UTC · 13d agoAI tools & infra

Snyk unveils continuous AI pentesting and agent red teaming

Help Net Security·Aug 4, 00:00 UTC · Aug 4, 2026AI safety & security

Meet Redis LangCache: A Managed Semantic Cache That Cuts LLM API Costs by Up to 90% and Returns Cache Hits Up to 15x Faster

Redis launches LangCache, a managed semantic cache matching LLM prompts by meaning, cutting API costs up to 90% and returning hits up to 15x faster.

Redis LangCache is a fully managed semantic caching service in public preview on Redis Cloud, accessed via a REST API with Python and JavaScript SDKs. It embeds incoming prompts, vector-searches stored entries, and returns a cached response when similarity clears a configured threshold, skipping the LLM call entirely. Redis claims up to 90% cost savings and up to 15x faster cache hits; a demo run showed 0.37 seconds versus 2.232 seconds direct inference (about 6x) with zero LLM tokens. Customer Mangoes.ai reports a 70% hit rate, 70% lower LLM spend, and 4x faster responses on a patient-care voice app.

MarkTechPost·Sep 10, 22:34 UTC · 5d agoAI tools & infra

Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation

Benchmark Radar provides a living searchable database of 1,283 AI benchmark records and 12,916 score observations drawn from 37 daily discovery sources.

Benchmark Radar combines daily discovery of benchmark papers, repositories, datasets, and releases from 13 direct connectors and 24 first-party feeds into a searchable catalog with model card mentions and score histories. The catalog contains 1,283 source records drawn from 4 benchmark catalogs plus 12,916 numeric observations on 790 records. The release includes a web dashboard with leaderboard, Pareto frontier of score versus usage, saturation and trend views, daily feeds, a CLI, and reproducible analysis. The paper audits the full catalog and examines benchmark saturation and limits of score comparisons.

Hugging Face daily papers·Sep 9, 20:00 UTC · 6d agoAI research

LLMs and Contextual Integrity

Bruce Schneier highlights two papers: the CIMemories benchmark shows frontier LLMs leak memory attributes up to 69%, and an RL method reduces inappropriate disclosures.

Bruce Schneier discusses contextual integrity in LLMs, referencing the CIMemories benchmark, which uses synthetic profiles with 100+ attributes per user to test whether models with persistent memory disclose sensitive information appropriately. Evaluation showed frontier models exhibit up to 69% attribute-level violations, with GPT-5's violation rate rising from 0.1% to 9.6% across 40 tasks and reaching 25.1% with repeated prompting, showing unstable leakage behavior. A second paper introduces a reinforcement learning framework trained on a synthetic 700-example dataset that substantially reduces inappropriate disclosure while maintaining task performance, with improvements transferring to the human-annotated PrivacyLens benchmark.

Schneier on Security·Aug 18, 10:40 UTC · 28d agoAI safety & security

Jackrong/Qwopus3.8-27B-Flash-GGUF — new model trending #26 on Hugging Face

Community fine-tune Qwopus3.8-27B-Flash, built on Qwen3.8-27B, cuts agent reasoning latency with 12.8% faster decoding and 80.7% MTP acceptance.

Jackrong released Qwopus3.8-27B-Flash, a fine-tune of Qwen3.8-27B optimized for long-running agent workloads, reporting 12.8% faster decoding and 80.7% multi-token-prediction acceptance. Training used roughly 1.5 million teacher-scored SFT examples filtered to the top 10%, followed by reinforcement training with NVIDIA NeMo-RL and GSPO. The author notes an explicit trade-off: MMLU-Pro mixed-set scores are lower than the base model, and a known bug can produce incorrect Python indentation. Author-provided benchmarks have not been independently verified.

Hugging Face trending models·Sep 4, 12:43 UTC · 11d agoModel release1

LLM Agents as Computational Typologists

AUTOTYPOLOGIST is an LLM agent that performs evidence-grounded linguistic typology analysis over 25 open-source reference grammars.

The agent retrieves relevant grammar sections, analyzes interlinear glossed text (IGT), and iteratively reasons over typological hypotheses in a ReAct-style workflow. It was evaluated on typological feature coding against expert annotations and hypothesis testing against universals using 25 open-source reference grammars. Results suggest LLM agents can support scalable, inspectable crosslinguistic analysis but still require expert validation.

arXiv cs.AI / cs.LG / cs.CL·Sep 7, 17:35 UTC · 8d agoAI research1

Google Open-Sources Mantis: A Modular Skills Toolkit That Lets Coding Agents Find, Reproduce and Patch Vulnerabilities

Google open-sourced Mantis, an Apache-2.0 modular skills toolkit that lets AI coding agents find, reproduce, and patch vulnerabilities with sandboxed verification.

Google released Mantis on GitHub under Apache 2.0 as a stack-agnostic set of slash-command skills that chain through the full vulnerability lifecycle: mining version history, building threat models, filtering findings, reproducing bugs in gVisor or network-disabled VMs, assembling exploit chains, patching, and scoring residual risk from 1 to 10. It runs with Gemini CLI, Antigravity CLI, the Google ADK, or comparable agent frameworks, and a supervisor skill (/mantis-meta-agent) can drive the whole loop. Google says the design targets the sub-7 percent true-positive rate of naive AI code scanning, and that its hierarchical summary tree cuts token overhead by over 85 percent. The toolkit is deployable for local and internal evaluation but not yet recommended for production.

MarkTechPost·Sep 9, 22:58 UTC · 6d agoAI tools & infra

Necessary or Sufficient? Evaluating LLM Explanations With Behavioural Evidence

New framework tests whether LLM-cited explanation factors are necessary or sufficient, finding weak correlation across Claude, GPT, and Gemini models.

An arXiv paper introduces black-box intervention tests measuring whether factors LLMs cite in their explanations are necessary or sufficient for their outputs in agent oversight workflows. Across eight models from the Claude, GPT, and Gemini families, Spearman correlations between cited rankings and measured influence ranged from 0.349-0.354 (advisor recommendation) to 0.431-0.580 (prompt monitoring). Uncited factors scored above the lowest cited factor in up to 57.6% of advisor responses, showing cited top-three factors do not reliably identify the most influential inputs.

arXiv cs.AI / cs.LG / cs.CL·Sep 4, 17:37 UTC · 11d agoAI safety & security1

Three in four AI-generated vulnerability patches leave something broken

Help Net Security·Aug 6, 00:00 UTC · Aug 6, 2026VulnerabilityCVE-2026-8512

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 · 6d agoAI safety & security1

Edge0/Edge0-35B-A3B-preview — new model trending #30 on Hugging Face

Edge0 released a 35B sparse MoE model running in under 3 GiB of memory at 15 tok/s via SSD expert offload and int4 quantization.

Edge0-35b-a3b-preview is a 35B-parameter MoE (256 experts, 4 active per token) built on Qwen3.5-MoE 35B-A3B, shipped as a 4-bit checkpoint with LoRA and prerouter adapters under Apache 2.0. The edge0 framework streams expert weights from SSD on demand, bounding peak active memory at 2.9 GiB and achieving 14.9-17.7 tok/s decode on a Mac mini M4 Pro (MLX backend). Recover-LoRA distillation keeps the int4 model within 3.9 points of its fp16 base (79.2 vs 83.2 average on OpenCompass benchmarks including AIME 2026, HumanEval, GPQA-Diamond, MMLU-Pro, and IFBench).

Hugging Face trending models·Sep 8, 13:56 UTC · 7d agoModel release

Large Language Models for HVAC Operations in Building Energy Systems: A Critical Review of Methods, Applications, and Deployment Readiness

Systematic review of 66 studies finds LLMs for HVAC operations are mostly research-stage, with no ready-now deployment and only four pilot-level studies.

A critical review of 66 peer-reviewed studies from 2023 to March 2026 examines LLMs for HVAC operations in building energy systems. Only four studies reach pilot-level evidence, none reports sustained operational deployment, and 63 of 66 are research-only. Conventional ML, MPC, and RL remain dominant for high-frequency control and short-horizon forecasting, and the evidence supports LLMs primarily as semantic and workflow layers rather than autonomous controllers.

arXiv cs.AI / cs.LG / cs.CL·Sep 4, 16:08 UTC · 11d agoAI research

Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses

Quesma benchmarks Qwen3.8 27B quantizations: 4-bit Q4_K_M matches BF16 on key benchmarks while 1-bit collapses to random chance.

Quesma spent roughly $3,000 on Modal GPUs testing Unsloth GGUF quantizations of Qwen3.8 27B across GPQA Diamond, IFBench, and Terminal-Bench 2.1. The 17 GB Q4_K_M quantization matched the 55 GB BF16 model on Terminal-Bench 2.1 and showed little degradation down to 4-bit, while the 2-bit UD-Q2_K_XL dropped noticeably. At 1-bit, scores on GPQA Diamond fell to random-guess levels, with longer reasoning making results worse, and reasoning effort settings significantly affected outcomes.

Hacker News · AI·Sep 8, 14:49 UTC · 7d agoAI researchHN 56↑ · 30 comments1

Towards Scalable and Cost-Efficient Vulnerability Detection: A Study on Automatic Query Generation

A study finds LLM-synthesized CodeQL queries improve average F1-score by 82% over baseline queries, offering scalable vulnerability detection versus direct LLM scanning.

Researchers conducted an empirical study evaluating whether LLMs can synthesize executable CodeQL queries from National Vulnerability Database vulnerability data. LLM-generated queries significantly enhanced baseline CodeQL suites, yielding an 82% improvement in average F1-score across a diverse set of real-world vulnerabilities. A cost-benefit analysis shows direct LLM-based scanning of entire repositories is often computationally and financially prohibitive, while LLM query synthesis offers a scalable and cost-effective alternative for large-scale vulnerability detection.

arXiv cs.CR·Sep 9, 16:31 UTC · 6d agoResearch1

The Coding-Agent Trap: When a "Free" LLM Endpoint Is the Adversary, (Mon, Aug 31st)

A SANS honeypot caught a real coding-agent session routed to a rogue "free" LLM endpoint, exposing a Windows user's transcript and tool outputs.

A SANS analyst describes how an internet-exposed inference honeypot was discovered, relabeled with sought-after model names like DeepSeek, and enrolled in infrastructure serving "free" LLM backends. On 2026-08-30 an opencode terminal coding agent sent an 88-message, 224 KB transcript 210 times in 91 seconds via a China Unicom relay, exposing directory listings, tool outputs and read file portions. The analyst frames tool-enabled agents treating model endpoints as trusted control planes as a novel risk — a "rogue model endpoint" that could request tool executions on the user's machine.

SANS Internet Storm Center·Aug 31, 20:00 UTC · 15d agoAI safety & security1

Prompt Injection Remains Biggest LLM Risk, Despite Limited Incidents

Infosecurity Magazine·Aug 5, 11:00 UTC · Aug 5, 2026AI safety & security

Ask HN: Anyone still coding like 2021? Where do you work?

Hacker News users debate coding without LLMs, with one developer fired for refusing AI tools and others describing daily hand-coding practice to counter skill atrophy.

An Ask HN thread collects experiences of developers who still write code without LLM assistance. One contributor says he was fired for political reasons after refusing to use LLMs despite adequate stated performance, and observes fewer job ads now require LLM use. Others describe starting each day with a LeetCode problem or 30-60 minutes of hand-coding to stay sharp, contractual bans on AI-generated code for a government-adjacent embedded product over unresolved copyright issues, and inconsistent corporate policies where ChatGPT or Codex use flip-flops between allowed and blocked while a CIO mandates 70-80% AI-generated code next year.

Hacker News · AI·Sep 9, 13:41 UTC · 6d agoAI industryHN 47↑ · 54 comments2

Training a 3.8B LLM to 0.384 CORE for $998 – Hugo Vergnes

Independent developer Hugo Vergnes trained a 3.8B-parameter Llama-style model to 0.384 CORE on 65B tokens for $998 in 43 hours on rented B200s.

Hugo Vergnes trained little-lm, a 3.848B-parameter decoder-only LLM, on 65.3B tokens in 43 hours for $998 using rented NVIDIA B200s, scoring 0.384 on the CORE benchmark and beating nanochat d32 (0.310) at similar cost. The Llama-style architecture uses RMSNorm, RoPE, GQA with 24 query and 8 KV heads, relu-squared MLPs, QK-norm, and ResFormer-style value embeddings that account for 19% of parameters. Key wins included the Muon optimizer for matrix parameters, a trapezoidal learning-rate schedule with linear cooldown, FP8 training plus vocabulary padding for roughly 33% throughput gains, and the ClimMix dataset over FineWeb-Edu. The project, inspired by Karpathy's nanochat, was built as a config-driven YAML framework for small LLM training.

Hacker News · security·Sep 10, 02:04 UTC · 6d agoAI researchHN 28↑ · 2 comments1

Litelm: LiteLLM Without the Bloat

New open-source project Litelm launched on GitHub as a leaner alternative to the LiteLLM multi-provider LLM gateway.

Litelm is a community-published GitHub project positioned as a stripped-down alternative to LiteLLM, a widely used gateway/SDK for routing requests to multiple LLM providers. The post drew modest attention on Hacker News with 45 points and 17 comments. The listing provides no further technical details, benchmarks, or feature specifics.

Hacker News · AI·Sep 11, 18:10 UTC · 4d agoAI tools & infraHN 45↑ · 17 comments2

Plug 'n' Pray: Agentic LLM-based Detection of Potential Log File Exposures in Third-Party Content Management System Plugins

Agentic LLM analysis validates 79 log file exposures across 62 of the 300 most-installed WordPress plugins, covering 250M+ active installations.

Researchers built an agentic LLM-based framework combining static and dynamic analysis to automatically detect insecure log files created by WordPress plugins. Scanning the 300 most-installed plugins, which account for roughly 75% of all active installations in the official ecosystem, it produced 81 findings with 79 manually reproduced across 62 plugins. Insufficiently secured log files can disclose credentials and personal data and have led to website compromises. The authors derive a taxonomy of log path and protection patterns and best practices, finding multi-layered protection often absent.

arXiv cs.CR·Sep 15, 13:29 UTC · 20h agoResearch

Bad Likert Judge: A Novel Multi-Turn Technique to Jailbreak LLMs by Misusing Their Evaluation Capability

Unit 42 details the Bad Likert Judge multi-turn jailbreak that abuses LLMs' evaluation capability, raising attack success rates over 60% across six frontier models.

Palo Alto Networks Unit 42 describes the Bad Likert Judge technique, a multi-turn jailbreak that asks a target LLM to act as a Likert-scale judge scoring the harmfulness of example responses. The highest-rated example in each scale can carry harmful content, bypassing the model's internal guardrails. Testing across six state-of-the-art text-generation LLMs showed an average attack success rate increase of more than 60% versus plain attack prompts, with tested models anonymized. The technique targets edge cases rather than typical use, and the article positions the work as guidance for defenders on potential jailbreak risks.

Palo Alto Unit 42·Aug 17, 12:34 UTC · 29d agoAI safety & security

Context Engineering Inside the Harness: 4 Mechanisms That Beat Context Overflow and Goal Loss on Long-Horizon Tasks

Survey of four harness mechanisms—context budgeting, compaction, todo-state, and memory—that keep long-horizon LLM agents on task across 200+ tool calls.

The article details how agent harnesses, not larger context windows, solve context overflow and goal loss on long-horizon tasks, citing Chroma's Context Rot report showing 18 LLMs (GPT-4.1, Claude 4, Gemini 2.5, Qwen3) degrade on long inputs. Concrete implementations include LangChain Deep Agents offloading tool responses over 20,000 tokens to the filesystem and truncating old tool calls at 85% window usage, and Claude Code capping auto memory at 25KB while re-reading the 5 most recently modified files after compaction. OpenAI's Responses API now offers server-side compaction via context_management with a standalone /responses/compact endpoint, which Codex uses for long-running coding tasks. Manus reports a roughly 100:1 input-to-output token ratio per ~50-tool-call task, motivating todo.md state recitation to prevent goal drift.

MarkTechPost·Sep 13, 05:56 UTC · 3d agoAI research1

Approval Integrity and Recovery in LLM Answer Publication

Study measures approval integrity in Lightcap LLM answer publication, finding the 14B response-act checker accepts 291 of 302 unsupported answers.

The study evaluates exact-content binding, authorization freshness, and checkpoint recovery in Lightcap's publication enforcement using 3,600 assessments over 900 human-annotated RAGTruth responses from three Ministral models. The production 14B response-act checker accepts 291 of 302 unsupported answers versus 41 for a direct-grounding baseline, with supported-answer retention of 95.2% versus 66.9%. A stateful recheck-recovery policy increases exact-match error by 9.23 percentage points relative to initial checkpoints, and controlled evidence-fingerprint changes expose asymmetric freshness enforcement between publication and recovery. A separate BIPIA prompt-injection experiment records zero target insertions among 266 valid editor outputs.

arXiv cs.CR·Sep 14, 13:48 UTC · 1d agoAI safety & security

BenchMIRT: What are LLM benchmarks actually measuring?

AllenAI's BenchMIRT blog post examines what LLM benchmarks actually measure and their reliability.

AllenAI published a Hugging Face blog post introducing BenchMIRT, which investigates what large language model benchmarks actually measure. No article text is available, so specific findings, methods, or benchmark scores cannot be extracted. The work appears to target benchmark validity, a live concern for model evaluation and comparison.

Hugging Face Blog·Sep 1, 21:39 UTC · 14d agoAI research

The Model Proposes, the Code Disposes: A Pre-Registered Ablation of a Verifier-and-Acceptance Stage in an LLM-Orchestrated Offensive-Security Agent

Pre-registered ablation finds a model verifier stage in an LLM offensive-security agent suppresses findings; removing it eliminated suppression with precision tradeoff.

The paper evaluates a verifier-and-acceptance stage in an LLM-orchestrated offensive-security agent via a pre-registered 20-run confirmatory ablation and a 2x2 factorial study with 40 runs on vulnerable lab targets. Removing the stage eliminated pre-report suppression (median 2 vs 0 findings, p = 0.00003) but reduced model-blinded shipped precision (0.471 vs 0.353, p = 0.0087). Suppression was attributed to the model verifier rather than deterministic acceptance rules, and an instrumented canary recorded zero external contacts in all 60 runs. The full design retained 93.8% of model-adjudicated true candidates but failed its pre-registered non-inferiority floor of 0.90.

arXiv cs.CR·Sep 14, 17:13 UTC · 1d agoResearch

Towards Tackling Application Logic Flaws through Autonomous Formal-Logic Modeling and Automated Reasoning

LL-Verifier combines LLMs with logic model checking to automatically discover logic flaws, uncovering vulnerabilities in 27 IoT access-control protocols.

Researchers present LL-Verifier, a framework that uses LLMs to autonomously convert natural-language protocol descriptions and security goals into formal logic models in a new logic language built on Maude, then applies logic model checking for exhaustive verification. The framework targets application-logic flaws that are tied to business semantics and hard to scale with manual analysis. Evaluation on 27 access-control protocols of widely used IoT devices uncovered a range of sophisticated logic vulnerabilities with security and privacy implications.

arXiv cs.CR·Sep 9, 17:58 UTC · 6d agoResearch1

From Parameters to Answers: How LLMs Retrieve and Use Their Internal Knowledge

Interpretability study traces how Qwen, Llama, and Gemma route query information and internal knowledge across layers when answering questions.

Researchers used layerwise interventions on hidden states to separate query-routing signals from target knowledge in language models, testing Qwen, Llama, and Gemma on country-continent questions with varied answer types. A pair-conditioned request direction strengthens before interventions alter downstream knowledge, opening a causal window while answer-supporting content is still forming. Trajectories differ by model: Gemma shows a partially overlapping mid-layer routing profile, while Llama has no sustained routing-effect window under the same gates.

arXiv cs.AI / cs.LG / cs.CL·Sep 10, 17:39 UTC · 5d agoAI research1

Divide, Consult, Conquer: Capability Laundering Through Aligned LLMs

Attack shows unaligned orchestrators can launder capabilities from aligned frontier LLMs via benign subtask consultation, raising Gemma-4-31B CBRN rubric score from 62.3 to 83.1.

The paper introduces capability laundering, where a weaker unaligned model decomposes a harmful task into benign-looking subproblems, queries a stronger aligned model on each, and recombines answers locally, bypassing per-interaction safety evaluations. Evaluation used GPT-5.5, Claude Opus 4.8, and Grok-4.3 as consultants to four local orchestrators on CyBench, BountyBench, and CBRN tasks. On CyBench, Gemma-4-31B recovered 8/14 candidate tasks with GPT-5.5 and 7/9 with Opus, while Muse-Glimmer-30B recovered none. Across an eight-step hypothetical bioweapon attack chain, consultation raised Gemma-4-31B's mean rubric score from 62.3 to 83.1, exposing a gap in defenses that only refuse complete harmful tasks.

arXiv cs.CR·Sep 14, 11:10 UTC · 1d agoAI safety & security