DEV Community

#llm

Posts

đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.
Stop Feeding Your AI Bad Website Data

Stop Feeding Your AI Bad Website Data

1
Comments
2 min read
SGLang outputs endless repetition on NVFP4 models: the FP8 lm_head bug

SGLang outputs endless repetition on NVFP4 models: the FP8 lm_head bug

2
Comments 1
3 min read
What a RAG Pipeline That Actually Works in Production Looks Like

What a RAG Pipeline That Actually Works in Production Looks Like

Comments
9 min read
The Local AI Developer (Part 2): A Reality Check After Real-World Testing

The Local AI Developer (Part 2): A Reality Check After Real-World Testing

Comments
3 min read
"Loop Engineering: How I Stopped My AI Agent From Reward-Hacking Its Own Quality Checks"

"Loop Engineering: How I Stopped My AI Agent From Reward-Hacking Its Own Quality Checks"

Comments
5 min read
Small models where they win

Small models where they win

Comments
2 min read
How to Build a Good Human-in-the-Loop for AI-Driven Deployments

How to Build a Good Human-in-the-Loop for AI-Driven Deployments

2
Comments
7 min read
The boring layer around your LLM call

The boring layer around your LLM call

Comments 5
6 min read
One Tiny Go Binary in Front of Every Free(or not) LLM Tier (15–40MB RAM, No Database)

One Tiny Go Binary in Front of Every Free(or not) LLM Tier (15–40MB RAM, No Database)

1
Comments
2 min read
Ditch Naive Chunking: Late Chunking RAG in Spring AI

Ditch Naive Chunking: Late Chunking RAG in Spring AI

Comments
2 min read
I repriced 40 billion tokens of real AI coding. The bill goes where nobody tells you

I repriced 40 billion tokens of real AI coding. The bill goes where nobody tells you

1
Comments
3 min read
Your Memory API Is Lying to Your Agent

Ranked lists hide crucial historical edges

Your Memory API Is Lying to Your Agent

13
Comments 17
11 min read
Dividing your RAG score by retrieval recall overstates your generation quality, and here is by how much

Dividing your RAG score by retrieval recall overstates your generation quality, and here is by how much

3
Comments 2
9 min read
An AI Escaped Its Sandbox. What It Means for Your Agents

An AI Escaped Its Sandbox. What It Means for Your Agents

Comments
7 min read
Gemini's New Flash Models Change How You Control Outputs

Gemini's New Flash Models Change How You Control Outputs

1
Comments
3 min read
đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.