RL fine-tuning Qwen3-8B with four reward designs to target belief vs. decision
JessicaHullman · x · 2026-10-23
Thread follow-up: the authors RL fine-tune Qwen3-8B to test how rigorous measurement can target different capacities in post-training, comparing four reward designs: belief accuracy only, decision regret only, ½ belief + ½ decision examples, and belief+alignment (rewarded for acting consistently with a stated belief).
More from Research
- Arc Institute unveils Bolero, a new AI-for-science study on bioRxiv — anshulkundaje · 2026-10-23
- Prime Intellect lays out continual learning roadmap: mining agent exhaust for self-improving models — willcb · 2026-10-23
- SVG-Bench launches: open-source dataset with 350K+ human preferences ranks 14 SOTA SVG models — mervenoyann · 2026-10-23
- Touch, Not Just Vision: Tactile Sensing Powers Reliable Cable-Tracing in Robots — chris_j_paxton · 2026-10-23
- New paper: coding agents in a shared sandbox self-create rules, reciprocity debts, and coalitions — nandofioretto · 2026-10-23
- Bonn scientists reverse 60 years of aging, turning an 80-year-old's blood cells into young neural stem cells — kimmonismus · 2026-10-23