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. 10-GEMMASCOPE-ATT-16K
    6. 6092
    Prev
    Next
    INDEX
    Explanations

    attends to the token marking the end of a sentence from the preceding clause of text that also includes punctuation

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k
    Configuration
    google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.10.attn.hook_z
    Hook Layer
    10
    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
    0:0.08
    1:0.12
    2:0.08
    3:0.39
    4:0.09
    5:0.06
    6:0.07
    7:0.07
    Negative Logits
    DAQ
    -0.35
     jurí
    -0.35
    quins
    -0.35
     blé
    -0.34
    leman
    -0.34
    CLICK
    -0.32
     congi
    -0.32
     Carlyle
    -0.32
    braucher
    -0.32
     dahin
    -0.31
    POSITIVE LOGITS
    AddTagHelper
    0.59
     nakalista
    0.58
    featureID
    0.58
     OMITBAD
    0.52
    +#+#
    0.50
     ostavi
    0.50
    InjectAttribute
    0.50
    rungsseite
    0.45
     snippetHide
    0.44
    verwijspagina
    0.44
    Activations Density 0.430%

    No Known Activations