© 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. 22-GEMMASCOPE-RES-131K
    7. 65251
    Prev
    Next
    INDEX
    Explanations

    mocking simulation data

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

    The neuron fires on numeric indicators—digits or count‐related tokens (e.g. numbers, “how many,” loop counters).

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

    code-related content, particularly variable assignments, function parameters, and technical documentation in programming contexts.

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

    programming and technical text, especially tokens from code snippets (keywords, parameters, numeric literals, and related jargon).

    oai_token-act-pair · gpt-5Triggered 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
     subclass
    -0.87
     compares
    -0.77
     Computing
    -0.76
     computing
    -0.75
     intolerant
    -0.75
     Violent
    -0.75
     vernacular
    -0.73
    subclass
    -0.73
    jelasan
    -0.73
     metrów
    -0.72
    POSITIVE LOGITS
     mock
    1.36
     simulated
    1.31
     simulate
    1.27
    mock
    1.25
     fictitious
    1.20
     simulation
    1.20
     fictional
    1.16
     simulating
    1.15
     fake
    1.13
    simulation
    1.13
    Activations Density 0.137%

    No Known Activations