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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 29195
    Prev
    Next
    INDEX
    Explanations

    instances of the word "ever" in conjunction with experiences or qualities being evaluated

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

    ever past participle

    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-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    parsedMessage
    -1.14
    tagHelperRunner
    -0.97
    <unused43>
    -0.94
    <unused79>
    -0.94
    <unused41>
    -0.94
    <unused8>
    -0.94
    <unused16>
    -0.93
    <unused23>
    -0.93
    <unused14>
    -0.93
    [@BOS@]
    -0.93
    POSITIVE LOGITS
    ,
    0.47
     
    0.44
    .
    0.41
     .
    0.39
     never
    0.37
     ever
    0.36
     created
    0.36
    <em>
    0.35
    Ever
    0.35
    ↵↵
    0.35
    Activations Density 0.011%

    No Known Activations