© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 67840
    Prev
    Next
    INDEX
    Explanations

    introducing a guess or question

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires on occurrences of the word “guess” (and similar question-prompt words like “which”), i.e. when the text introduces a guessing or rhetorical question.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Když
    -0.96
    utro
    -0.96
    ό
    -0.94
     precisam
    -0.88
     régiment
    -0.88
     pipette
    -0.87
    Questão
    -0.86
    blich
    -0.85
    Obrigado
    -0.85
    ₚ
    -0.84
    POSITIVE LOGITS
     guess
    1.33
     Guess
    1.14
     likely
    1.05
    Guess
    1.02
     bandou
    1.02
    guess
    1.01
     guessing
    0.97
     jamás
    0.95
     naturally
    0.93
     predic
    0.93
    Activations Density 0.025%

    No Known Activations