© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 11521
    Prev
    Next
    INDEX
    Explanations

    trends and lists

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

    The neuron reliably spikes on the very first token of each new text segment (e.g. paragraph or “<start>” section), effectively detecting segment/paragraph boundaries.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    line breaks or paragraph boundaries in low-quality machine-generated or spam text.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    contractions marked by apostrophes (’ or ') in English words and phrases.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Ди
    0.81
    Environment
    0.80
    {
    0.78
     Cybersecurity
    0.78
    的前
    0.77
    Cyber
    0.76
     Seag
    0.76
    Conte
    0.75
    -(\
    0.75
     Evergreen
    0.73
    POSITIVE LOGITS
    agogues
    0.86
    alerts
    0.77
     tendencias
    0.75
    ec
    0.74
    aham
    0.73
    bed
    0.73
    e
    0.73
    ాం
    0.72
     dudas
    0.72
    үкт
    0.72
    Activations Density 0.000%

    No Known Activations