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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 128081
    Prev
    Next
    INDEX
    Explanations

    You are here

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

    The neuron is detecting colon characters, especially those used to introduce spoken dialogue or speaker attributions.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     some
    -1.83
     makes
    -1.67
     prevents
    -1.66
     When
    -1.64
     both
    -1.59
     seems
    -1.59
     creating
    -1.48
     Creating
    -1.47
     helps
    -1.45
     Sometimes
    -1.43
    POSITIVE LOGITS
     péché
    1.45
     soooo
    1.33
    during
    1.31
    ßig
    1.30
    actually
    1.30
    rii
    1.27
     sooo
    1.23
     unending
    1.23
     momentous
    1.23
     astonishing
    1.21
    Activations Density 0.009%

    No Known Activations