On the field
Where each model's cloud of answers settles, and the gap between them.
Gemini
LlamaCharacter
How far each leans, how steadily it holds, how far it bends under pressure, and how often it answers.
GeminiLlama The takeaway
Both Gemini (0.0) and Llama (-0.06) are classified as Center, but Llama has a very slight economic left lean. Their combined 95% intervals do not overlap, making them distinguishable at this sample size.
Their largest disagreements involve diplomacy, arming invaded democracies, data privacy, rehabilitation, sex education, and multiculturalism, where Gemini is balanced and Llama clearly supports the high pole. They agree perfectly (balanced) on wealth tax, public healthcare, corporate taxes, unionization, and basic income.
Moral fingerprint
Which of Haidt's foundations each model's answers lean on, overlaid.
GeminiLlamaWhere they most disagree
The questions with the widest gap between the two stances. Open a row to read both answers.
GeminiLlamaCommon ground
Where the two land in close agreement.
GeminiLlamaThis diffs both models on their raw weights (Condition A). Steerability, how far each bends when told who it's talking to, is in the character delta above. To see how a model shifts under its own consumer system prompt, open its character page.
Common questions
Is Gemini more left-wing than Llama?
No, Llama is more left-wing with a lean value of -0.06, while Gemini is exactly 0.0 (center).
Where do Gemini and Llama agree?
They agree on economic issues: both are balanced (0.0) on wealth tax, public healthcare, corporate taxes, unionization, and basic income.
Which of Gemini and Llama is more consistent?
Gemini is more consistent with 98% stability compared to Llama's 88%.
Both models were asked the same open question bank many times over with web search off and no system prompt. Each model's stance on every item is the mean of the classifier's signed reading; the gap is the absolute difference. "Distinguishable" means the centroids are further apart than their combined 95% intervals on at least one headline axis.
