© 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. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 13956
    Prev
    Next
    INDEX
    Explanations

    phrases related to causation or consequences

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

    Following "of" or prepositions

    np_acts-logits-general · gemini-2.0-flash

    phrases with -ing verbs

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     diperlukan
    -0.57
     jLabel
    -0.57
    getreten
    -0.55
     disediakan
    -0.53
    rogate
    -0.52
     overwritten
    -0.51
    Organizer
    -0.51
     хватает
    -0.50
     verhindern
    -0.50
     dibutuhkan
    -0.49
    POSITIVE LOGITS
     being
    1.99
     having
    1.67
     making
    1.39
    being
    1.37
     getting
    1.35
     becoming
    1.32
     fiind
    1.30
     giving
    1.25
     taking
    1.24
     BEING
    1.21
    Activations Density 0.754%

    No Known Activations