» Tag
llm
473 postsFine-Tuning Qwen2-VL on AMD MI300X: The ROCm Pitfalls Docs Skip
A development log on fine-tuning Qwen2-VL for blockchain graph classification on AMD MI300X, detailing real ROCm setup and debugging friction.
How Airbnb Cut LLM Evaluation Time From Weeks to a Day
Airbnb's four-layer LLM evaluation stack uses deterministic caching and micro LoRA adapters to cut iteration time from weeks to a single day.
Qwisp: MoE expert-streaming engine runs Qwen3.6-35B-A3B on 8GB Macs
Qwisp streams MoE experts from flash to run the 35B-parameter Qwen3.6-A3B model on 8GB Macs, with bit-exact lossless decoding and raw-Metal speed.
CommitBrief — AI code reviews, right in your terminal
A provider-agnostic, local-first CLI that reviews your staged changes, a historic range, or a whole GitHub pull request. Zero telemetry, no server. Free and open source.
commitbrief.comContext Warp Drive: Deterministic, Zero-LLM Context Compaction
Context Warp Drive folds agent context deterministically with no LLM calls, cutting costs up to 72% while keeping 94% fact retention.
An Open-Source, Assessment-First Course: doerkit and rubric-bench
doerkit and rubric-bench: an open-source course using an LLM as rubric judge, plus a reusable regression-testing framework for any LLM judge.
When Agent Tools Return Empty: Who Decides What It Means
Why empty results from agent tool calls hide three distinct failure states, and how typed states, evidence, and retry semantics fix the gap.
Testing LLMs Like Software: A Promptfoo Deep Dive for QA
A production-level guide to testing LLM systems with Promptfoo: versioned evals, deterministic checks, judge-model bias, and provider benchmarking.
Memory Scarcity and Open Models Are Reshaping the AI Industry
A new paper models how the DRAM/HBM price crisis, open-weight models, and inference efficiency gains will reshape the AI industry through 2026-2030.
LLM-as-a-Verifier Turns Verification Into a New Scaling Axis
New research scales LLM verification without extra training, introducing continuous scoring that hits state-of-the-art on SWE-Bench, Terminal-Bench and more.
Probelock: A Lockfile for LLM Tool-Calling Capabilities
Probelock measures an LLM's tool-calling capabilities with deterministic tests and halts CI when a model, quantization, or runtime swap causes a regression.