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-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 30281
    Prev
    Next
    INDEX
    Explanations

    mentions of oil companies and their operations

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

    Names of companies and corporations

    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
    UnusedPrivate
    -0.36
     error
    -0.31
     habits
    -0.30
    生活中
    -0.30
     []:
    -0.29
     speech
    -0.29
     settings
    -0.28
     intro
    -0.28
     Yii
    -0.28
     Lifetime
    -0.28
    POSITIVE LOGITS
    rungsseite
    0.74
     subsidiaries
    0.71
     informée
    0.71
     companies
    0.68
    RTLI
    0.67
     corporations
    0.65
    companies
    0.63
     perusahaan
    0.63
     subcontractors
    0.62
     subsidi
    0.61
    Activations Density 0.039%

    No Known Activations