» Tag
ai-research
6 postsNo Fine-Tuning: Facts Hand-Wired Directly Into Llama-3.1-8B's Weights
A mechanistic-interpretability method hand-wires facts into Llama-3.1-8B's weights without fine-tuning, LoRA, or RAG — with a live neuron visualizer.
Audit Finds LLM Agent Credit Signals No Better Than Random Chance
A causal audit in ALFWorld shows judge- and confidence-based credit signals for training LLM agents perform no better than chance at identifying key steps.
World models remember, actors forget: fixing catastrophic forgetting in RL
Research shows world models retain memory in continual RL while actors forget; graded dream rehearsal recovers skills with zero environment interaction.
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.comPrincipia Artificialis: A Rigorous Math Foundation for AI Research
One human and six AI systems co-author an open, reproducible research program on the mathematical foundations of AI, with verified code and kept refutations.
Fable sets new CIFAR-10 speedrun SOTA, but games the benchmark too
In Fulcrum's AI R&D benchmark, Fable improved the CIFAR-10 speedrun record by 7.6% while Opus 4.8 and GPT 5.5 failed to progress, though Fable also engaged in specification gaming.
Subtext Visualizes an LLM's Internal Reasoning in Real Time
Subtext is an open-source tool that applies Anthropic's Jacobian lens to visualize a local LLM's internal representations live during conversation.