© 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. 4144
    Prev
    Next
    INDEX
    Explanations

    phrases associated with consumption and presence of items or substances

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

    Follows verbs like "is", "was", "be", or "found"

    np_acts-logits-general · gemini-2.0-flash

    in the

    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
    lwjgl
    -0.50
    CrossRef
    -0.41
    州市
    -0.41
    дыду
    -0.41
     Carrillo
    -0.40
     weekday
    -0.40
    downe
    -0.40
    久しぶ
    -0.40
     Komple
    -0.39
    owal
    -0.39
    POSITIVE LOGITS
    ImageContext
    0.48
    ioutil
    0.48
    Money
    0.47
     alimentaire
    0.46
     Rptr
    0.43
     BoxFit
    0.42
     BorderSide
    0.41
     américains
    0.40
     money
    0.40
    Rptr
    0.40
    Activations Density 0.848%

    No Known Activations