© 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
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 76431
    Prev
    Next
    INDEX
    Explanations

    topics related to race, reparations, and systemic injustices

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

    complaining about

    np_acts-logits-general · gemini-2.5-flash-lite

    calls to action or prescriptive policy proposals—especially arguments about obligations or payments (like reparations)—framed with collective “we/they should/need to” or imperative language.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
     Económica
    -0.54
    nsito
    -0.45
     Secretaría
    -0.42
     eşofman
    -0.42
     desmotivaciones
    -0.42
     accesibles
    -0.40
     atmósfera
    -0.40
     creatividad
    -0.40
     lavadora
    -0.40
     bañera
    -0.40
    POSITIVE LOGITS
    UnusedPrivate
    0.46
    prepareStatement
    0.42
     cid
    0.42
    classnames
    0.41
     fhould
    0.41
    pare
    0.40
     ſhould
    0.40
    pied
    0.40
    Sage
    0.39
    OrBuilder
    0.39
    Activations Density 0.090%

    No Known Activations