© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-2B
    7. Residual Stream - 16k
    9. 22-GEMMASCOPE-RES-16K
    11. 6554
    Prev
    Next
    INDEX
    Explanations

    expressions of desire or intention to avoid negative consequences or associations

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

    Following "to" or a negative verb

    np_acts-logits-general · gemini-2.0-flash

    avoiding negative outcomes

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.hook_resid_post
    Hook Layer
    22
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Modific
    -0.55
     doInBackground
    -0.54
     endeavouring
    -0.53
    </thead>
    -0.53
     endeavoured
    -0.53
    migrationBuilder
    -0.53
    Vanjske
    -0.52
    RegistryLite
    -0.52
     Ausnahme
    -0.51
     '\\;'
    -0.51
    POSITIVE LOGITS
     offend
    0.91
     risk
    0.79
     offending
    0.78
     upsetting
    0.72
     embarrassment
    0.72
     spoilers
    0.69
     disappoint
    0.69
     embarrass
    0.68
     lose
    0.67
     admitting
    0.67
    Activations Density 0.308%

    No Known Activations