cd/sources/hugging-face-blog· home› sources› Hugging Face Blog
cat /sources/hugging-face-blog.feed | wc -l → 1408

Hugging Face Blog

articles 1408 domain huggingface.co → page 7/71 feed RSS
21:33
2026-11-10
huggingface.co
ai-agents

343 calibration matches did not establish semantic conformance

An audit of a synthetic multi-agent project's terminal-event extractor found that version v1.1.2 matched all 343 disclosed calibration expectations but returned incorrect results on 12 of 58 deliberat…

14:15
2026-11-10
huggingface.co
large-language-models

Open-Source LLM Evaluation and Annotation QA Benchmark

A developer released an open-source Python benchmark on Hugging Face for LLM response evaluation and annotation quality assurance, built around weighted scoring across four dimensions, Pydantic-based …

00:00
2026-11-10
huggingface.co
artificial-intelligence

Decision Models in Hugging Face

Hugging Face published a blog post introducing "decision models," a model class that returns a probability distribution over predefined answer choices instead of generating text token by token, with t…

11:55
2026-11-09
huggingface.co
large-language-models

Vetting an open model before deployment

A small team replacing a paid API with an open Hub model found that public leaderboard scores did not predict production problems, according to a Hugging Face forum post by user DHEIVER. Testing on re…

04:38
2026-11-09
huggingface.co
ai-safety

Tuning guardrails without blocking real users

A fintech team that added PII redaction, topic restrictions and output moderation to a financial FAQ assistant found after launch that users were being refused on ordinary questions such as how intere…

16:36
2026-11-08
huggingface.co
ai-safety

What red teaming means for AI systems

Red teaming AI systems requires repeated testing of prompt injection, jailbreaks, system-prompt extraction and harmful-output requests, because models, prompts and retrieval data change and an attack …

13:05
2026-11-08
huggingface.co
large-language-models

When translating prompts saves tokens but loses meaning

A browser-local experiment called English-First Prompt Lab, built by Hugging Face user vitalipavlovsk with AI assistance, found that translating prompts into English reduced input tokens but often cha…

15:37
2026-11-06
huggingface.co
ai-agents

Harness adapters: Safe RL for efficient agentic development

Hugging Face released a set of "harness adapters" — one LoRA adapter per model and agent harness, trained with reinforcement learning inside unmodified pi, opencode and hermes harnesses — with every t…

← prev page 7 / 71 next →