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. 29434
    Prev
    Next
    INDEX
    Explanations

    references to judges or panels involved in evaluations or selections

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

    judges and juries

    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
     ujednoznacz
    -0.75
    PreferredItem
    -0.72
    tvguidetime
    -0.64
     незавершена
    -0.60
     الرياضيه
    -0.60
    цездатний
    -0.60
     الحره
    -0.59
    oa̍t
    -0.57
    adpleegd
    -0.57
     informée
    -0.56
    POSITIVE LOGITS
     judges
    0.61
     judging
    0.60
     judge
    0.56
     jury
    0.54
     Jury
    0.50
     jurado
    0.47
     jueces
    0.47
     Judges
    0.43
     juges
    0.41
     jurors
    0.41
    Activations Density 0.013%

    No Known Activations