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-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 992
    Prev
    Next
    INDEX
    Explanations

    instances of the word "listen" and its variations, indicating a focus on auditory engagement

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

    "listen" or variations of it

    np_acts-logits-general · gemini-2.0-flash

    listen to

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    protoimpl
    -0.92
    rungsseite
    -0.92
     autorytatywna
    -0.89
    ConstraintMaker
    -0.85
     Chwiliwch
    -0.84
    醐
    -0.79
     Abp
    -0.79
    posedge
    -0.78
    bootstrapcdn
    -0.77
    Datuak
    -0.75
    POSITIVE LOGITS
     listen
    1.09
     Listening
    1.08
     listening
    1.05
     listened
    1.03
     LISTEN
    1.02
     listeners
    1.02
     Listen
    1.01
    Listening
    1.00
    Listen
    0.99
     listens
    0.97
    Activations Density 0.045%

    No Known Activations