© 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-2-9B
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 1935
    Prev
    Next
    INDEX
    Explanations

    components and concepts related to writing and journalism skills

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

    details about writing, argumentation, and analytical skills in educational contexts.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    language about formal writing and its structure—argumentation, evidence, document sections/organization, and editorial or review directives.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    structural or organizational elements in formal written texts, such as section markers, numbered points, and discourse connectives that guide the reader through arguments or document organization.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
    adone
    -0.39
    //
    -0.39
     Lets
    -0.37
    +#+
    -0.37
    беріга
    -0.36
    ichthys
    -0.35
    operability
    -0.35
    predicates
    -0.35
     debu
    -0.35
    fillStyle
    -0.34
    POSITIVE LOGITS
     purpoſe
    0.65
     fubject
    0.55
    AndEndTag
    0.54
     ſtate
    0.53
     Anſ
    0.52
     narr
    0.51
     Reſ
    0.50
     preſent
    0.50
     Inſ
    0.49
     Grecs
    0.49
    Activations Density 0.167%

    No Known Activations