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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 85662
    Prev
    Next
    INDEX
    Explanations

    punctuation and conjunctions that connect clauses in complex sentences

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

    This followed by punctuation

    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-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    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
     AssemblyCompany
    -0.67
    جغرافيا
    -0.65
    msgTypes
    -0.64
    FormTagHelper
    -0.64
     ModelExpression
    -0.63
    RegressionTest
    -0.62
     مرئيه
    -0.62
    Personensuche
    -0.60
    AnchorTagHelper
    -0.59
    pushFollow
    -0.59
    POSITIVE LOGITS
     aimed
    0.33
     event
    0.33
     ditu
    0.31
    (/\
    0.31
     inspired
    0.31
     inspirada
    0.31
    Eins
    0.30
     safety
    0.30
     bolso
    0.30
     cél
    0.29
    Activations Density 0.031%

    No Known Activations