© 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. 70957
    Prev
    Next
    INDEX
    Explanations

    what they differ is

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

    The neuron detects evaluative or contrastive cue words (e.g. “differ,” “what,” “matter,” “more,” “surprising,” “however”) that introduce questions or emphasize comparisons.

    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
     to
    -1.06
    }$.
    -1.02
    eably
    -0.98
    龵
    -0.96
    addGap
    -0.94
    democracy
    -0.94
    naphthal
    -0.92
    ].
    -0.92
     paillettes
    -0.92
    阏
    -0.91
    POSITIVE LOGITS
    However
    1.22
    卻
    1.17
    ,
    1.02
     are
    1.02
     however
    0.96
     is
    0.95
     invece
    0.91
     Equally
    0.90
    яс
    0.89
     Biaya
    0.88
    Activations Density 0.021%

    No Known Activations