© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 16192
    Prev
    Next
    INDEX
    Explanations

    phrases and expressions that highlight critical issues and often overlooked aspects in various topics

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

    Something overlooked or underrated

    np_acts-logits-general · gemini-2.0-flash

    overlooked, unnoticed, neglected

    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
    DebuggerNonUser
    -0.96
    MemoryWarning
    -0.92
     initComponents
    -0.86
     ProtoMessage
    -0.79
    makeConstraints
    -0.75
     NSCoder
    -0.74
    antMatchers
    -0.70
    __(/*!
    -0.69
    Sklici
    -0.69
    SharedDtor
    -0.66
    POSITIVE LOGITS
     overlooked
    1.92
     forgotten
    1.72
     unnoticed
    1.57
     neglected
    1.56
     underrated
    1.50
     hidden
    1.48
    forgotten
    1.47
     ignored
    1.36
     obscure
    1.33
    Hidden
    1.31
    Activations Density 0.394%

    No Known Activations