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. Attention Out - 16k
    5. 10-GEMMASCOPE-ATT-16K
    6. 5528
    Prev
    Next
    INDEX
    Explanations

    attends to tokens related to concepts or items marked as "purpose," "day," "and," or "clinical" from subsequent tokens with a similar context

    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
    1:0.13
    2:0.08
    3:0.09
    4:0.36
    5:0.06
    6:0.07
    7:0.09
    Negative Logits
    IntoConstraints
    -0.35
     BoxDecoration
    -0.31
    onAttach
    -0.25
     themſelves
    -0.25
    StandardCharsets
    -0.25
     canst
    -0.24
    estacks
    -0.24
    sidemargin
    -0.23
     himſelf
    -0.23
     ἐκ
    -0.23
    POSITIVE LOGITS
    rungsseite
    0.35
    Datuak
    0.34
    msgTypes
    0.32
    Personendaten
    0.31
    ImageContext
    0.31
     nahilalakip
    0.28
     odb
    0.28
    aarrggbb
    0.28
     ujednoznacz
    0.27
    StructEnd
    0.27
    Activations Density 0.618%

    No Known Activations