The Archive

SpecValidator detects defective task descriptions in LLM code generation via lightweight parameter-efficient classifier.

Green Shielding: user-centric evaluation framework measuring LLM sensitivity to benign input variations via CUE criteria.

Persona Collapse in multi-agent LLM simulations: agents converge to homogeneous behavior despite distinct profiles; framework measures Coverage, Uniformity, Complexity.

SciCrafter: Minecraft benchmark evaluating agents' discovery-to-application loop via parameterized redstone circuit tasks.

Contextual Linear Activation Steering (CLAS) dynamically adapts steering strength per token, outperforming fixed-strength methods across 11 benchmarks.

Noise-Based Spectral Embedding (NBSE): physics-informed feature selection via Nishimori temperature without greedy search.

ProHist-Bench: LLM historical reasoning benchmark grounded in Chinese Imperial Examination system, evaluating evidentiary reasoning over lexical knowledge.

Informational Viability Principle for autonomous AI agent governance: runtime monitoring and restriction via unobserved risk bounds without code changes.

Benchmarks pathology foundation models on breast cancer survival prediction from histopathology images using standardized evaluation.

BMW case study: adapts LLMs to generate multi-file domain-specific language code for Xtext-based DSL with Java/TypeScript output.

Reddit discussion on peer review feedback distinguishing research papers from technical reports; meta-commentary on publication standards.

..that too while the task was in b/w - Claude is disappointing me these days... For context: I'm just repromtping something that was generated - it was simply to edit the content

Luce DFlash: open-source speculative decoding for Qwen3.6-27B achieves 1.98x throughput on RTX 3090 via GGUF port.

Mathematical formulation of convolutional neural networks via presheaves/copresheaves on topological spaces to analyze architectural limitations.

Measures sycophancy (agreement bias) in LLMs deployed for agentic financial tasks under user contradiction.

Evaluates whether LLMs track Turkish evidential morphology conditioned on source trustworthiness in cloze tasks.

Proposes belief-space MPC for control of linear systems with bilinear observations where control affects state estimation quality.

Introduces DySIB method to learn low-dimensional state representations from high-dimensional time-series via information bottleneck principle.

Proposes agent-native research artifacts to eliminate storytelling and engineering taxes in scientific papers for AI agent consumption.

AgentWard: defense-in-depth lifecycle security architecture for autonomous AI agents spanning initialization through execution.

OpenAI and Microsoft's partnership-turned-situationship just got even less committed. And a clause about artificial general intelligence, which has for years dictated the future of their deal, has officially been dropped. On Monday morning, Microsoft announced a handful of big changes to its long-standing OpenAI deal. Microsoft will remain OpenAI's "primary cloud partner, and OpenAI products will ship first on Azure, unless Microsoft cannot and chooses not to support the necessary capabilities." But OpenAI can "now serve all its products to customers across any cloud provider." That lets Open...

DepthKV: layer-dependent KV cache pruning reduces memory for long-context LLM inference by pruning low-attention tokens per layer.

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. In February, I picked up a flyer at an anti-AI march in London. I can’t say for sure whether or not its writers meant to riff on South Park’s underpants gnomes. But…

K-MetBench: expert-level multimodal benchmark for Korean weather forecasting LLMs, exposing gaps in visual reasoning and domain knowledge across 55 models.

Skye's new AI app attracted investors before it even launched — a sign of interest in a more AI-aware iPhone.

Functional Task Networks: neocortex-inspired parameter-isolation method for continual learning using self-organizing binary masks over expert networks.

Production case study: integrating small LMs (Gemma 4E2B, Qwen3 0.6B) into Android app reveals memory, latency, and quantization challenges for on-device deployment.

Computational screening of 5M photoactive PARP1 inhibitor candidates using ML and atomistic simulation for light-activated cancer drugs.

User reports GBNF grammar optimizations for Qwen 3.6 35B and 27B models improving coding task performance in llama.cpp.

Senator Josh Hawley questions Helen Toner (former OpenAI) on AI labor displacement and existential risk, signaling congressional scrutiny of frontier AI development.