Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    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. 885
    Prev
    Next
    INDEX
    Explanations

    expressions of emotional volatility and aggression

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

    Negative mood and behavior

    np_acts-logits-general · gemini-2.0-flash

    negative emotions and behavior

    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
     continúas
    -0.55
    Viitteet
    -0.53
    PerformLayout
    -0.48
    StringCopy
    -0.48
    omitempty
    -0.48
     doInBackground
    -0.47
    mány
    -0.47
    GOTREF
    -0.46
     Solidar
    -0.46
    Ос
    -0.45
    POSITIVE LOGITS
     behavior
    1.34
     temper
    1.32
     behaviour
    1.32
    behavior
    1.23
     Behavior
    1.21
    behaviour
    1.18
     Temper
    1.18
     behave
    1.17
     Behaviour
    1.16
     comportement
    1.16
    Activations Density 0.340%

    No Known Activations