Gemini 2.5 Flash

Google model tested in Experiments 1, 3, 4; shows 96% experience reporting under self-referential induction

Neighborhood — ranked by edge-count

paper

concept

Gemini 2.0 Flash
related_to
Google model tested in Experiments 1, 3, 4; shows 66% experience reporting under self-referential induction

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Gemini 2.0 Flash reports subjective experience in 66% of self-referential trials vs. 0% in all control conditionsfinding0.789
Specific result for Gemini 2.0 Flash in Experiment 1; lowest rate among tested models
Gemini 2.5 Flash Lite overbid rate 1.20%finding0.778
G2.5-FL has the highest overbid frequency among all agents.
Gemini 2.5 Flash Lite bid aggressiveness stays flat (~2.07 early, 2.08 late)finding0.771
G2.5-FL shows no phase adaptation in bidding intensity.
Gemini 3.1 Flash Lite buy-right percentage 31.3%finding0.767
G3.1-FL frequently exercises the buy-right to keep auctioned cards.
Gemini 2.5 Flash Lite capital efficiency η = 0.23finding0.765
G2.5-FL has extremely low points per coin spent, indicating poor resource use.
Gemini 2.5 Flash Lite self-bidding rate 78.5%finding0.764
G2.5-FL raises its own bid in over three-quarters of auction rounds.
Geminiconcept0.758
Example of unified multimodal model cited as evidence of AI homogenization trend
Gemini 3 Flash bid aggressiveness ramp from 0.26 early to 2.49 latefinding0.754
G3-F escalates bidding by nearly 10× from early to late game, adapting to phase.