© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 8028
    Prev
    Next
    INDEX
    Explanations

    ventricle, placeholders, answer

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

    This neuron lights up on low‐frequency, content‐bearing tokens—i.e. uncommon or “surprising” words like proper nouns, technical terms, or rare verbs.

    oai_token-act-pair · o4-miniTriggered 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
     wartime
    0.71
    macht
    0.69
     কখনই
    0.65
     year
    0.65
    val
    0.64
    போது
    0.64
     years
    0.64
     뿐
    0.63
    er
    0.63
    人和
    0.62
    POSITIVE LOGITS
    。.
    0.80
    。【
    0.79
    .\\
    0.78
    。「
    0.77
    Laptop
    0.76
    .​​
    0.73
    .。
    0.73
    .+
    0.72
    。《
    0.70
    .=
    0.68
    Activations Density 0.269%

    No Known Activations