Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 103084
    Prev
    Next
    INDEX
    Explanations

    legal and judicial language concerning rights and permissible actions

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

    Modal auxiliaries and qualifying conjunctions in formal written discourse, particularly in legal and academic contexts where precise expression of permissions, possibilities, and limitations is required.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    Priority
    -0.40
    Ӕ
    -0.38
    siyon
    -0.35
    PRIORITY
    -0.34
     Priority
    -0.32
     full
    -0.32
     womb
    -0.31
     peł
    -0.31
     volumetric
    -0.31
    <0x8C>
    -0.30
    POSITIVE LOGITS
     <<<<<<<<<<<<<<
    0.64
    RenderAtEndOf
    0.59
    parsedMessage
    0.58
     mergeFrom
    0.54
    CloseOperation
    0.52
    BeginContext
    0.51
    HasAnnotation
    0.50
     Audiodateien
    0.49
    styleType
    0.49
     naturali
    0.49
    Activations Density 0.198%

    No Known Activations