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. 5763
    Prev
    Next
    INDEX
    Explanations

    programming-related method calls

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

    user input marker

    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
    исленность
    -1.17
     would
    -0.36
     estancia
    -0.35
    _
    -0.35
     An
    -0.35
     Dave
    -0.35
     
    -0.35
     G
    -0.34
     térm
    -0.34
     $
    -0.34
    POSITIVE LOGITS
    原始内容存档于
    5.88
    原始内容存档
    2.06
    原始内容
    1.62
     оригіналу
    1.56
     跳转至
    1.53
    Архівовано
    1.34
    存于互联网档案馆
    1.17
     مؤرشف
    1.13
     محفوظة
    1.12
     مشين
    1.06
    Activations Density 0.000%

    No Known Activations