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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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. 102396
    Prev
    Next
    INDEX
    Explanations

    terms related to constructive and polite interactions

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

    supervision, anonymous, constructive, controlled, reasonable, consenting, proper, resistant

    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
    IVEREF
    -0.42
     beträgt
    -0.35
     particulières
    -0.33
    Espèce
    -0.32
     contributed
    -0.32
     auprès
    -0.32
     hề
    -0.32
     Tenemos
    -0.31
     particulière
    -0.31
    Cześć
    -0.31
    POSITIVE LOGITS
    bootstrapcdn
    0.81
    ftagPool
    0.61
    ьаж
    0.59
    GEBURTSDATUM
    0.56
     respectful
    0.56
     Chwiliwch
    0.54
    elemField
    0.53
     ब्रेकडाउन
    0.52
    ########.
    0.52
    Jeografia
    0.52
    Activations Density 0.040%

    No Known Activations