© 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-2-9B-IT
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 14987
    Prev
    Next
    INDEX
    Explanations

    instances of deception or trickery

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

    Deception, trickery, or being misled

    np_acts-logits-general · gemini-2.0-flash

    trickery and deception

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
     înal
    -0.47
     gesteld
    -0.45
     betrekking
    -0.43
     uitges
    -0.42
     betreft
    -0.41
    forte
    -0.41
     Linton
    -0.41
     behov
    -0.41
     Tennyson
    -0.40
    ennia
    -0.40
    POSITIVE LOGITS
     trick
    0.98
     tricks
    0.90
     Trick
    0.79
     Tricks
    0.78
    trick
    0.76
     sneaky
    0.75
     tricked
    0.75
     trucos
    0.71
    Trick
    0.71
     scam
    0.69
    Activations Density 0.462%

    No Known Activations