© 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. 22-GEMMASCOPE-RES-131K
    4. 26
    Prev
    Next
    INDEX
    Explanations

    modal verbs followed by verbs or nouns

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

    The neuron lights up on common English function words and discourse markers (e.g. “this,” “may,” “be,” “because,” “it,” “will,” “apply,” “sometimes,” “does”) that typically occur in explanatory or instructional sentences.

    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
     aren
    -1.02
     momente
    -0.99
     frase
    -0.99
     azonban
    -0.96
     didn
    -0.96
    ::$_
    -0.93
     if
    -0.93
     but
    -0.90
     energie
    -0.88
     however
    -0.84
    POSITIVE LOGITS
     wtedy
    1.20
     znaczy
    1.05
     później
    1.04
     einerseits
    0.99
     HAST
    0.99
     offenbar
    0.96
    没关系
    0.94
    ङ
    0.93
     zwykle
    0.93
     ambulances
    0.93
    Activations Density 0.051%

    No Known Activations