© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 103795
    Prev
    Next
    INDEX
    Explanations

    states or events

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

    The neuron strongly activates on longer, multi-syllabic content words (e.g. “arises,” “discovered,” “protective,” “interlocked,” “canceled”), effectively detecting token length.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     ensuing
    -1.20
     —
    -1.20
    --
    -1.16
     its
    -1.16
     --
    -1.10
     decorativa
    -1.02
     it
    -1.02
     consistente
    -1.02
     pronta
    -0.97
     isn
    -0.96
    POSITIVE LOGITS
     in
    1.15
    vá
    1.05
    eeeeeeee
    0.98
     内衣
    0.96
    ränken
    0.93
    بير
    0.92
    كري
    0.91
     Probl
    0.91
    ම
    0.90
    %).
    0.89
    Activations Density 0.022%

    No Known Activations