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. 32-GEMMASCOPE-RES-131K
    6. 129987
    Prev
    Next
    INDEX
    Explanations

    underscores in various contexts, possibly indicating elements in structured data or code

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

    __u8, __be16, __be32

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_32/width_131k/average_l0_88
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.32.hook_resid_post
    Hook Layer
    32
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     کاش
    -0.36
    Sem
    -0.30
    全
    -0.30
    ↵↵
    -0.29
     of
    -0.29
    2
    -0.29
     MA
    -0.28
    like
    -0.28
    ...
    -0.28
    All
    -0.28
    POSITIVE LOGITS
     müſſen
    0.79
     dieſen
    0.79
     queſta
    0.75
    <pad>
    0.73
    <unused23>
    0.73
    <unused8>
    0.73
    <unused68>
    0.73
     dieſem
    0.73
    <unused16>
    0.72
    <unused52>
    0.72
    Activations Density 0.024%

    No Known Activations