Skip to content
Navigation menu
Search
Powered by Algolia
Search
Log in
Create account
DEV Community
Close
#
llm
Follow
Hide
Posts
Left menu
đź‘‹
Sign in
for the ability to sort posts by
relevant
,
latest
, or
top
.
Right menu
The KV Cache Is the Bottleneck: A 2026 Field Guide to Attention Variants
ai maya
ai maya
ai maya
Follow
Aug 9
The KV Cache Is the Bottleneck: A 2026 Field Guide to Attention Variants
#
ai
#
llm
#
machinelearning
#
transformers
1
 reaction
Comments
1
 comment
4 min read
Prompt Injection Isn't Fixed by a Filter. It's Fixed by Architecture
Eugen Ullrich
Eugen Ullrich
Eugen Ullrich
Follow
Jul 21
Prompt Injection Isn't Fixed by a Filter. It's Fixed by Architecture
#
ai
#
architecture
#
llm
#
security
Comments
Add Comment
22 min read
Attention Sinks: Why Streaming LLMs Break When You Evict Token 0
jidonglab
jidonglab
jidonglab
Follow
Jul 17
Attention Sinks: Why Streaming LLMs Break When You Evict Token 0
#
deeplearning
#
llm
#
machinelearning
#
nlp
Comments
Add Comment
6 min read
Your Pinned OpenAI Models Stop Working Next Week
albe_sf
albe_sf
albe_sf
Follow
Jul 17
Your Pinned OpenAI Models Stop Working Next Week
#
ai
#
llm
#
machinelearning
#
devtools
Comments
Add Comment
3 min read
Tips on Handling Files in LLMs
iwan rustiawan
iwan rustiawan
iwan rustiawan
Follow
Jul 17
Tips on Handling Files in LLMs
#
ai
#
backend
#
llm
#
tutorial
Comments
Add Comment
2 min read
SDABench: A New Benchmark for Evaluating LLMs in Scientific Discovery
Pneumetron
Pneumetron
Pneumetron
Follow
Jul 17
SDABench: A New Benchmark for Evaluating LLMs in Scientific Discovery
#
llm
#
scientificdiscovery
#
benchmark
#
airesearch
Comments
Add Comment
4 min read
Length Penalties in LLMs: Shorter Chains of Thought, Hidden Influences
Pneumetron
Pneumetron
Pneumetron
Follow
Jul 17
Length Penalties in LLMs: Shorter Chains of Thought, Hidden Influences
#
llm
#
chainofthought
#
reinforcementlearning
#
interpretability
Comments
Add Comment
3 min read
Deep Interaction: A Novel Approach to Correcting LLM Reasoning Errors
Pneumetron
Pneumetron
Pneumetron
Follow
Jul 17
Deep Interaction: A Novel Approach to Correcting LLM Reasoning Errors
#
llm
#
ai
#
humanaiinteraction
#
chainofthought
Comments
Add Comment
3 min read
LangChain4j and Spring AI: The Plumbing to make your Java Apps talk to LLMs
shayesta
shayesta
shayesta
Follow
Jul 17
LangChain4j and Spring AI: The Plumbing to make your Java Apps talk to LLMs
#
java
#
ai
#
springboot
#
llm
Comments
Add Comment
8 min read
Is there a standard way to use one API key across GPT and cheaper models from other vendors in the same app?
pkuchenhua
pkuchenhua
pkuchenhua
Follow
Jul 17
Is there a standard way to use one API key across GPT and cheaper models from other vendors in the same app?
#
llm
#
api
#
typescript
#
saas
Comments
1
 comment
6 min read
Your cache_read_input_tokens is zero. Here is what silently did it.
Jackson Ly
Jackson Ly
Jackson Ly
Follow
Jul 30
Your cache_read_input_tokens is zero. Here is what silently did it.
#
llm
#
ai
#
performance
#
api
1
 reaction
Comments
2
 comments
5 min read
Getting Started with Kimi K3: API Setup, Code Examples, and First Impressions
Khuram Youraj
Khuram Youraj
Khuram Youraj
Follow
Jul 17
Getting Started with Kimi K3: API Setup, Code Examples, and First Impressions
#
ai
#
api
#
llm
#
tutorial
Comments
Add Comment
4 min read
Agentic Browser: ~98% fewer tokens than HTML for LLM web agents (Python + MCP)
TOPHET
TOPHET
TOPHET
Follow
Jul 17
Agentic Browser: ~98% fewer tokens than HTML for LLM web agents (Python + MCP)
#
python
#
ai
#
llm
#
mcp
Comments
1
 comment
1 min read
The Watermark Is the Real Boundary Between Training and Serving
AI Explore
AI Explore
AI Explore
Follow
Jul 21
The Watermark Is the Real Boundary Between Training and Serving
#
ai
#
aie
#
llm
#
programming
1
 reaction
Comments
Add Comment
4 min read
Human-in-the-Loop Checkpoints in LangGraph
sekera-radim
sekera-radim
sekera-radim
Follow
Jul 17
Human-in-the-Loop Checkpoints in LangGraph
#
ai
#
llm
#
opensource
#
tutorial
Comments
Add Comment
3 min read
đź‘‹
Sign in
for the ability to sort posts by
relevant
,
latest
, or
top
.
We're a place where coders share, stay up-to-date and grow their careers.
Log in
Create account