© 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. 84603
    Prev
    Next
    INDEX
    Explanations

    inaccessible spaces

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

    The neuron is looking for words that denote small hidden recesses or the act of getting into them (e.g. “nooks,” “crannies,” “into,” “down,” “deep”).

    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
    sohn
    -0.84
    ]));
    -0.82
    ynthesis
    -0.81
     itali
    -0.75
    áh
    -0.74
     глав
    -0.74
    ponym
    -0.73
     contagios
    -0.72
     Cho
    -0.71
    rody
    -0.71
    POSITIVE LOGITS
     crevices
    1.97
     nook
    1.69
     spaces
    1.57
     places
    1.56
     corners
    1.55
     inaccessible
    1.41
     hidden
    1.41
     Spaces
    1.38
    hidden
    1.35
     трудно
    1.34
    Activations Density 0.047%

    No Known Activations