Why shared review checklists can hurt: variance reduction must exceed added bias
JessicaHullman · x · 2026-10-30
Theoretical motivation of the work: a shared checklist applied to papers by AI with high accuracy and shown to reviewers can increase consistency, but a shared view correlates errors — reviewers may fixate on it over their private information. For a checklist to work, the variance it reduces must exceed the bias it adds.
More from Research
- Commonroom beta: chat with neuron-steered LLM personas and hunt emergent personalities — repligate · 2026-10-30
- Elorian launches visual reasoning model Glimpse alongside DecoyCount benchmark — AndrewDai · 2026-10-30
- Instruction Capacity paper quantifies rule-following limits: GPT-4o mini holds ~37 rules at 95% accuracy — Hennadii Kutomanov · 2026-10-30
- Guidelabs launches Causal Diffusion Engine, adding interpretability to diffusion LLMs with no latency cost — juliusadml · 2026-10-30
- Alice & Bob says RSA-2048 falls in 3.14 days with 23,000 cat qubits — jedisct1 · 2026-10-30
- FlexGlue's 0.6 px average keypoint shift matches SIFT's subpixel refinement threshold — ducha_aiki · 2026-10-30