META-IR Uses Meta-Learning for Adaptive Safety Alignment Against Evolving Jailbreaks
META-IR is a meta-introspective reasoning framework for adapting LLM safety policies to ev…
ZICQ curates daily AI news and deep dives on models, open-source infrastructure, and developer field reports.
Submit an articleCurrent filter: Tag: #Surrogate Modeling Clear filters
META-IR is a meta-introspective reasoning framework for adapting LLM safety policies to ev…
LLMs & Foundation Models· 2026-11-10
AYO (AI Youth Open Science) is an open-source research platform for studying how children …
Compute & Architecture· 2026-11-10
LoRazor is a multi-LoRA serving framework designed for shared-backbone deployments. It all…
LLMs & Foundation Models· 2026-11-10
A study evaluated LLM-based annotation of crisis severity in veteran-related social media …
Research & Papers· 2026-11-10
A study spanning three model families, three benchmarks, and nine model-benchmark combinat…
LLMs & Foundation Models· 2026-11-10
CulturalCrosswords is a multilingual benchmark built from crowdsourced cultural riddles in…
Research & Papers· 2026-11-10
Researchers use Qwen3.5-27B as a model organism for studying evaluation awareness. Through…
Agentic· 2026-11-10
Clinical agents can solve routine tasks yet still make unsafe decisions when tools return …
Research & Papers· 2026-11-10
Multimodal large language models can memorize and reproduce sensitive information from bot…
LLMs & Foundation Models· 2026-11-10
DramaGrow introduces a regime-aware benchmark and audit protocol for LLM judges evaluating…
LLMs & Foundation Models· 2026-11-10
A benchmark study evaluates decoder-only LLMs—including Llama, Mistral, Gemma, and Phi 4—o…
LLMs & Foundation Models· 2026-11-10
CoMixKV addresses the KV-cache memory bottleneck that appears when multiple LLM agents rem…