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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-2B
    7. Residual Stream - 16k
    9. 18-GEMMASCOPE-RES-16K
    11. 15423
    Prev
    Next
    INDEX
    Explanations

    references to dangerous or aggressive actions or threats

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

    following prepositions and conjunctions

    np_acts-logits-general · gemini-2.0-flash

    disrupting and sabotaging

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Administrativna
    -0.60
    ViewFeatures
    -0.56
    Архівовано
    -0.53
     tamen
    -0.53
     optim
    -0.52
     ottim
    -0.51
     optimizer
    -0.49
     otim
    -0.49
     forgiving
    -0.49
    SPJ
    -0.49
    POSITIVE LOGITS
     disruption
    0.86
     destabili
    0.85
     disrupt
    0.81
     disrupting
    0.78
     sabotage
    0.77
     forcing
    0.77
     cripple
    0.74
     paral
    0.74
     intimidate
    0.73
     paraly
    0.72
    Activations Density 0.489%

    No Known Activations