{"slug": "small-open-weight-language-models-for-crowd-decision-making-capabilities-and-of", "title": "Small Open-Weight Language Models for Crowd Decision-Making: Capabilities and Limits of Output-Level Fusion and Linguistic Aggregation", "summary": "A study of five open-weight language models with 8-36B total parameters, run on a single consumer GPU against a hosted reference, found that repeating and fusing model outputs stabilizes crowd decision-making winners but does not eliminate sensitivity to which reviewers are sampled. On 474 TripAdvisor reviews of four London restaurants queried ten times each, redrawing the reviewers kept the winner on 47-81% of replicas versus 86-100% for redrawing the runs, and with simultaneous intervals over every pair of alternatives no system on either the London corpus or a 600-review Madrid corpus of twelve restaurants named a single winner. The three larger open-weight models matched the hosted reference's London winner and at least matched its per-review correlation with reviewers' star ratings, leading the authors to recommend reporting candidate sets with margins rather than forced rankings.", "body_md": "arXiv:2611.17589v1 Announce Type: new \nAbstract: Small open-weight language models can run crowd decision-making (CDM) on an institution's own infrastructure, which keeps the checkpoint, the inference settings and the opinions under its control. Their judgements vary across repeated queries; whether repeating and aggregating them, often through the 2-tuple linguistic representation, settles the decision rather than the labels is untested. Five open-weight models of 8-36B total parameters on one consumer GPU and a hosted reference classify 474 TripAdvisor reviews of four London restaurants ten times each; three open-weight models repeat the protocol on 600 reviews of twelve Madrid restaurants. Continuous evaluations preserve resolution: snapping each review leaves 5 distinct values against 21-65, and snapping at every level at most two collective values. Sensitivity to the participants persists once run-to-run variability is reduced: in London, redrawing the reviewers keeps the winner on 47-81% of replicas against 86-100% for redrawing the runs. With simultaneous intervals over every pair of alternatives under crowd resampling, no system on either corpus names a single winner: in London every system keeps all four restaurants as candidates, and in Madrid the candidate sets hold five to eight of the twelve. Extra runs and output-level fusion stabilize the winner but do not remove that sensitivity, and the fusion does not beat its best member. The three larger open-weight models name the hosted reference's London winner and at least match its per-review correlation with the reviewers' stars, so small open-weight models can support local CDM, reported as candidate sets with margins rather than forced rankings.", "url": "https://wpnews.pro/news/small-open-weight-language-models-for-crowd-decision-making-capabilities-and-of", "canonical_source": "https://www.machinebrief.com/news/small-open-weight-language-models-for-crowd-decision-making-psy6", "published_at": "2026-11-16 05:00:00+00:00", "updated_at": "2026-11-16 05:24:45.772991+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "machine-learning", "artificial-intelligence"], "entities": ["TripAdvisor", "London", "Madrid"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/small-open-weight-language-models-for-crowd-decision-making-capabilities-and-of", "markdown": "https://wpnews.pro/news/small-open-weight-language-models-for-crowd-decision-making-capabilities-and-of.md", "text": "https://wpnews.pro/news/small-open-weight-language-models-for-crowd-decision-making-capabilities-and-of.txt", "jsonld": "https://wpnews.pro/news/small-open-weight-language-models-for-crowd-decision-making-capabilities-and-of.jsonld"}}