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. 13838
    Prev
    Next
    INDEX
    Explanations

    phrases and verbs related to the action of filling or filling out

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

    "fill" or "filling"

    np_acts-logits-general · gemini-2.0-flash

    fill out or in

    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
    RegressionTest
    -0.77
    LookAnd
    -0.67
    AndroidJUnit
    -0.59
    UTERS
    -0.57
    databind
    -0.57
    verick
    -0.56
     RequestMethod
    -0.56
     LOGGER
    -0.56
    acon
    -0.55
    ogada
    -0.55
    POSITIVE LOGITS
     Fill
    1.09
    fill
    1.09
    Fill
    1.08
     fill
    1.07
    filling
    1.05
     Filling
    1.05
     filling
    1.04
    Filling
    1.01
    FILL
    0.99
    filled
    0.98
    Activations Density 0.131%

    No Known Activations