© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    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. 594
    Prev
    Next
    INDEX
    Explanations

    references to progress or continuity

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

    "Far" as an adverb or part of a phrase

    np_acts-logits-general · gemini-2.0-flash

    up to this point

    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
    writerow
    -0.31
     voulu
    -0.31
    isie
    -0.30
     tristesse
    -0.29
    gonic
    -0.28
    setToolTipText
    -0.28
     sociedad
    -0.28
     debido
    -0.28
     Gees
    -0.27
     due
    -0.27
    POSITIVE LOGITS
    fjspx
    0.87
    为止
    0.84
    dafx
    0.78
     sejauh
    0.77
     bisher
    0.75
    setVerticalGroup
    0.71
    featureID
    0.68
    CppMethod
    0.68
     protoimpl
    0.67
     bislang
    0.67
    Activations Density 0.115%

    No Known Activations