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-2-9B
    4. Attention Out - 16k
    5. 0-GEMMASCOPE-ATT-16K
    6. 1283
    Prev
    Next
    INDEX
    Explanations

    attends to instances of the digit "1" from sequences containing other digits, focusing on a specific pattern or relation to "1"

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61
    Prompts (Dashboard)
    16,384 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
    0:0.01
    1:0.01
    2:0.46
    3:0.12
    4:0.03
    5:0.03
    6:0.01
    7:0.08
    8:0.01
    9:0.01
    10:0.01
    11:0.03
    12:0.04
    13:0.04
    14:0.02
    15:0.01
    Negative Logits
    ]');
    -0.43
    ]");
    -0.40
    ]);
    -0.40
    ]);
    
    -0.40
    ]):
    -0.39
    ])
    -0.38
    ]).
    -0.38
    ]]);
    -0.38
    ].)
    -0.37
    ]
    -0.36
    POSITIVE LOGITS
     Italij
    0.33
     orejas
    0.32
     dioses
    0.32
     Dichtung
    0.31
     morire
    0.31
    rungsseite
    0.30
     époux
    0.29
     Divulgação
    0.29
     cielos
    0.29
     titolata
    0.29
    Activations Density 0.429%

    No Known Activations