© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 15-GEMMASCOPE-RES-16K
    6. 12005
    Prev
    Next
    INDEX
    Explanations

    instances of the word "never" and its variations, indicating a focus on negation or denial

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Hook Layer
    15
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    GenerationType
    -0.82
    stdc
    -0.81
     fibrillation
    -0.74
    thanol
    -0.72
    :]:
    -0.72
    AOC
    -0.70
     Muffins
    -0.69
     MDR
    -0.69
     Rioja
    -0.68
     UVB
    -0.67
    POSITIVE LOGITS
     Never
    1.18
     never
    1.17
    NEVER
    1.15
     NEVER
    1.14
    never
    1.13
     EVER
    1.12
    Never
    1.09
     Ever
    1.08
    ever
    1.07
     ever
    1.01
    Activations Density 0.054%

    No Known Activations