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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Attention Out - 16k
    5. 0-GEMMASCOPE-ATT-16K
    6. 346
    Prev
    Next
    INDEX
    Explanations

    attends to the token "avoid" from corresponding tokens representing negative outcomes or challenges

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k
    Configuration
    google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.attn.hook_z
    Hook Layer
    0
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.70
    1:0.01
    2:0.01
    3:0.01
    4:0.16
    5:0.01
    6:0.01
    7:0.04
    Negative Logits
    iffion
    -0.58
     saites
    -0.57
     كومونز
    -0.52
    etheless
    -0.51
    ^(@)
    -0.50
     myſelf
    -0.50
    Personensuche
    -0.49
    ſelves
    -0.49
    adeloupe
    -0.49
     itſelf
    -0.48
    POSITIVE LOGITS
    j
    0.43
    L
    0.42
    I
    0.40
    ac
    0.40
    it
    0.37
    R
    0.37
    is
    0.37
    id
    0.36
    sp
    0.36
    eng
    0.36
    Activations Density 0.009%

    No Known Activations