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
    4. Residual Stream - 16k
    5. 21-GEMMASCOPE-RES-16K
    6. 15639
    Prev
    Next
    INDEX
    Explanations

    specific starting sequences or tokens at the beginning of a segment

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.hook_resid_post
    Hook Layer
    21
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ロウィン
    -0.58
    ⇿
    -0.56
    久しぶ
    -0.55
    essoal
    -0.54
    ConstraintSet
    -0.53
    SupportActionBar
    -0.52
     ContentValues
    -0.51
    YSQL
    -0.50
    batore
    -0.50
    joker
    -0.49
    POSITIVE LOGITS
    toxicity
    0.80
    :✨
    0.40
    Diweddarwch
    0.40
    \{\\
    0.40
    addPreferredGap
    0.39
     Taktlose
    0.38
    ViewFeatures
    0.35
     pinulongan
    0.35
    hoeddwyd
    0.35
     Meksiku
    0.34
    Activations Density 0.037%

    No Known Activations