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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 15145
    Prev
    Next
    INDEX
    Explanations

    variations and details related to procedural processes and assessments

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

    preceding verbs like "maintain," "enhance," "increase"

    np_acts-logits-general · gemini-2.0-flash

    maintain, enhance, increasing, complete

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
     and
    -0.31
    ;
    -0.27
    t
    -0.26
    y
    -0.25
     however
    -0.24
    ,
    -0.23
     but
    -0.22
    2
    -0.21
    快
    -0.20
    <eos>
    -0.20
    POSITIVE LOGITS
    ſicht
    1.05
    ſſung
    1.03
     メンテナ
    1.00
     zwiſchen
    0.99
    <unused3>
    0.99
    [@BOS@]
    0.99
    <unused41>
    0.99
    <unused47>
    0.98
    <unused28>
    0.98
    <unused8>
    0.98
    Activations Density 0.515%

    No Known Activations