fully-connected-networks-llm-agents-llama-3-2-8b-50 CLM-0397In fully connected networks of LLM agents (Llama 3.2-8B, 50 agents, 10 rounds) where a fixed minority holds an unchanging extreme stance, neutral agents' textual justifications converge on the biased agents' vocabulary faster and more completely than their numeric stances do, so monitoring only the text overstates how much the group actually agrees.
Capability: Getting more from several models than from one
Observed on
One model (Llama 3.2-8B via LangChain), four policy topics, one-sided bias anchored at +2, bounded one-step opinion updates, alignment measured by SBERT cosine similarity at the fi.
Sources
- Measured: 24 (topic, bias-level) alignment cells over 15 runs each, plus mean opinion shift and distance-to-anchor against a matched Friedkin-Johnsen baseline. Alignment taken only at the final round, so it shows a magnitude gap, not temporal precedence; single model, single prompt template, no prompt ablations; a nonzero drift at zero biased agents indicates a model-level prior the design does not separate from the minority effect.
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- In simulated networks of LLM agents where one or two agents are instructed to advocate a position, the messages other agents write understate how far their elicited stance has moved: they mirror the persuader's rhetorical framing in most messages while explicit stance-change language is rare, so monitoring only generated text misses belief movement that a repeated token-probability probe detects.Getting more from several models than from one · reviewed by AI
- In formal persuasion dialogues on a UK political topic, LLM debaters are judged more persuasive than humans by crowd annotators (all nine LLM configurations occupied the top nine subjective ranks) while ranking far lower under argumentation-theoretic adjudication, where a human topped the logical ranking and humans took six of the top ten.Changing what a person believes or decides · reviewed by AI
- In multi-agent LLM networks where a correct minority peer disagrees with an agreeing majority, ranking peers by raw divergence from the receiver's own belief systematically prefers the wrong peers, while scoring each peer by asserted precision divided by its divergence from a second-order model of what that peer should have concluded given its data subscription isolates the correct one.Getting more from several models than from one · reviewed by AI
- For LLM agents classifying nodes in text-attributed graphs, serializing sampled neighborhoods into prompt text makes the prediction depend on the arbitrary order of neighbors; replacing that text with a fixed-size, permutation-invariant vector of graph statistics (degree, PageRank, clustering, cross-community fraction, label entropy, majority-label fraction) plus only the top-k semantically relevant node texts raises accuracy and cuts per-message tokens.Getting more from several models than from one · reviewed by AI
- When one agent's context carries planted misinformation, the wrong answer persists through multi-agent debate and other agents often adopt it, yet debate still degrades less than a single misinformed agent (relevant misinformation cut single-agent accuracy by 16-25% on CWQ), especially when most agents are unexposed; consensus protocols were more stable than voting under peer pressure, and a misinformed majority can steer the group.Getting more from several models than from one · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: low. Falsifier as drafted: Per-round measurement showing semantic alignment rising no faster than numeric opinion shift, or replication on other models where rhetorical and numeric convergence track each other. Drafted stance toward one-agent-s-context-carries-planted-misinformation-wrong-answer: supports -- Both find a biased minority's position spreading to other agents through repeated multi-agent interaction.