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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 9103
    Prev
    Next
    INDEX
    Explanations

    references to magical abilities and their utility in combat or escape scenarios

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

    Magic spells and abilities

    np_acts-logits-general · gemini-2.0-flash

    game specific terms

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    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
    原始内容存档于
    -0.45
    Wre
    -0.43
    computers
    -0.42
    そうそう
    -0.39
    Supplement
    -0.39
     Normdatei
    -0.39
     diritti
    -0.38
    出版年
    -0.38
    SqlClient
    -0.38
     Flüssigkeit
    -0.38
    POSITIVE LOGITS
    MigrationBuilder
    0.56
    AnchorStyles
    0.48
    ParallelGroup
    0.44
    pexpr
    0.42
     BoxDecoration
    0.42
     HasFactory
    0.41
     AssemblyVersion
    0.39
    GraphicsUnit
    0.39
     tatuagens
    0.39
    ỡng
    0.38
    Activations Density 0.454%

    No Known Activations