Program-as-Weights compiles task descriptions into reusable local CPU functions
yuntiandeng · x · 2026-09-21
Following a discussion on cross-task generalization, the author details Program-as-Weights: a larger model is trained to turn English function descriptions directly into LoRA adapters for a shared 0.6B model.
At deployment you describe a new task, compile it once, and reuse the resulting function locally on CPU—the standard compiler emits the adapter without per-task finetuning. Code, weights, and the paper are all public.
More from Research
- MoME: Context-Aware Mixture-of-Memory Embeddings Outperform Value Embedding Baselines at Iso-FLOPs — vector-institute · 2026-09-21
- FRAUDSkill Boosts Audio Anti-Fraud Macro-F1 to 73.5% With Frozen Weights — PPSUCTeleantifraudCommunity · 2026-09-21
- TeleAntiFraud 2.0: Chinese Audio Fraud Benchmark Shows F1 Drops to 0.65 on Near-Domain Negatives — PPSUCTeleantifraudCommunity · 2026-09-21
- Fitting a performance cone with bootstrapped models to test if AI leaderboard rankings actually hold — PTenigma · 2026-09-21
- AI writing detectors flagged her lab vision post; she asks what we should actually measure — furongh · 2026-09-21
- Human-AI Collaboration Settles Major Open Problem in Multi-Winner Voting Theory — xuanalogue · 2026-09-21