© 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. 107326
    Prev
    Next
    INDEX
    Explanations

    pyramid and pyramidal neurons

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

    references to pyramidal structures or shapes, whether in architecture, computer programming, or neuroscience.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron selectively activates on words containing the “pyramid” (or “pyramidal,” “pupiform,” etc.) root, i.e. terms referencing pyramid‐shaped structures or the prefix “pyrami-.”

    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
    גש
    -0.91
     purview
    -0.82
     arrogant
    -0.80
    bilecek
    -0.77
    անո
    -0.75
    picks
    -0.74
     particolari
    -0.73
     professionali
    -0.73
    Snapshots
    -0.73
    edor
    -0.72
    POSITIVE LOGITS
     scheme
    1.04
     schemes
    1.04
     SCHEME
    1.00
    scheme
    0.91
     Scheme
    0.90
     texts
    0.86
     Texts
    0.83
    Scheme
    0.83
     Claire
    0.80
    鱒
    0.80
    Activations Density 0.027%

    No Known Activations