© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-2B
    7. Residual Stream - 65k
    9. 20-GEMMASCOPE-RES-65K
    11. 267
    Prev
    Next
    INDEX
    Explanations

    verbs and phrases related to attachment and holding on

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    expressions of hope and attachment

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    cling to

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Gambas
    -0.77
    Personendaten
    -0.71
    twimg
    -0.67
    `.`
    -0.61
    NullCheck
    -0.60
     endwhile
    -0.59
    vantaged
    -0.58
    bisk
    -0.58
    setVerticalGroup
    -0.58
    dillas
    -0.57
    POSITIVE LOGITS
     clings
    0.84
     clinging
    0.82
     dear
    0.82
     tightly
    0.73
     clung
    0.73
     tight
    0.71
     hope
    0.69
     cling
    0.68
     hold
    0.65
     dearly
    0.64
    Activations Density 0.033%

    No Known Activations