W&B Weave (@weave

W&B Weave

572 posts

W&B Weave

@weave_wb

A lightweight toolkit for tracking and evaluating LLM applications, built by @wandb for AI developers!

Land of GPUs

Joined October 2024

Pinned
W&B Weave
@weave_wb
Sep 17, 2025
Your RL run just spiked at step 89! But, do you know why? We’re fixing that. Today we’re launching W&B Weave Traces to give you a step by step look into your agent’s decisions. This is the first drop from our fresh new integration with @OpenPipeAI. More RL magic is incoming.
00:00
291K
W&B Weave
@weave_wb
Aug 22, 2025
BIG LAUNCH: @deepseek_ai's V3.1 is now live on W&B Inference! One model, two modes: toggle between high-speed 'Non-Think' & deep 'Think'. Priced at just $0.55/$1.65 per 1M tokens, it’s a game-changer for building intelligent agents. Want $50 in free credits? Details below.
425K
W&B Weave
@weave_wb
Aug 4, 2025
🚨 Building LLM apps in production? You need guardrails that do more than just block bad content, they should also help you measure and improve output quality. Here’s how W&B Weave lets you do both with unified scorers below!
00:00
395K
W&B Weave
@weave_wb
Mar 18, 2025
GenAI moves fast. W&B Weave helps your AI workflows move faster. Build, debug, monitor, and scale the next generation of AI apps—so your team can innovate, iterate, and shape the future together. This is just the beginning.
00:00
224K
W&B Weave
@weave_wb
Aug 29, 2025
W&B Inference is now on @openrouter! You can use @AIatMeta's Llama 3.3 70B Instruct today, with more models rolling out in the coming weeks.
187K
W&B Weave
@weave_wb
Aug 8, 2025
Ever graded a TTS model and felt the “good/bad” score missed the real story? Rubric evaluation turns fuzzy opinions into weighted, reproducible numbers. We teamed up with the @encord_team and pitted two TTS models against 20 criteria, here’s the play-by-play. 👇
203K
W&B Weave
@weave_wb
May 29, 2025
🚀 Our latest free course “AI engineering: Agents” drops June 2! Learn to ship production AI agents via tool-chains + memory, multi-agent orchestration, evals and much more. Built with @OpenAIDevs and taught by @ilanbigio + @ash0ts. Grab your spot ⤵️ wandb.me/AIagentsx
305K
W&B Weave
@weave_wb
Jun 26, 2025
The AI talent war is making AI labs look more like sports teams, but @DavidCahn6 from @sequoia thinks founders are still the most valuable players.
00:00
354K
W&B Weave
@weave_wb
Apr 18, 2025
Here’s your no-fluff roundup of what dropped this week—from GPT-4.1’s million-token brain to Gemini's response 👀 🧵
125K
W&B Weave
@weave_wb
May 9, 2025
Relying on public benchmarks to choose a coding LLM? Rookie mistake. We ran Gemini 2.5 Pro, o4-mini, and Claude 3.7 Sonnet through 30 fresh CodeContests problems using W&B Weave to track real-world performance. Here’s what we found! 🧵
120K
W&B Weave
@weave_wb
Aug 26, 2025
Building AI that answers with confidence requires two things: fresh context and transparent execution. Static agents miss both. We paired @tavilyai (real-time web) with W&B Weave (tracing, evals, ops) to ship research agents that are accurate, current and trustworthy.
58K
W&B Weave
@weave_wb
Apr 29, 2025
AI agent tracing shouldn’t slow you down. W&B Weave now auto-logs MCP traces with one line of code—zero setup, no clutter. Coming soon: plug-and-play for Google’s A2A protocol. Build, ship, repeat. 😤
42K
W&B Weave
@weave_wb
Jun 9, 2025
Fine-tuning LLMs for function-calling is quietly becoming a power move in the agentic AI era. If you’re building autonomous systems that think before acting, this thread breaks down when it’s worth it and exactly how to do it right. 🧵
53K
W&B Weave
@weave_wb
Aug 13, 2025
🚀 New in W&B Weave: Assets Prompts, Datasets and Scorers now live in one view and tie cleanly into applications and evaluations. This makes for a cleaner nav so teams can reuse components, track provenance, and keep AI evals reproducible by default.
00:00
132K