© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 41977
    Prev
    Next
    INDEX
    Explanations

    quotation marks and apostrophes, indicating dialogue or quoted speech

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

    delimiters and structures

    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
    了嗎
    -0.47
    ')],
    -0.38
     rootView
    -0.38
    ]:
    -0.36
     siguientes
    -0.35
     genomen
    -0.34
    ()];
    -0.34
    `,
    -0.33
    >';
    -0.33
    verket
    -0.33
    POSITIVE LOGITS
     "
    1.18
    ("
    1.08
     '
    1.02
    []{"
    0.98
    !("
    0.98
     “
    0.96
    ('
    0.92
     مشين
    0.90
     ("
    0.88
     ["
    0.88
    Activations Density 0.011%

    No Known Activations