© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Initial Gemma-2B SAEs
    5. Gemma-2B
    7. Residual Stream
    9. 0-RES-JB
    11. 2676
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "tower" with varying emphasis

    oai_token-act-pair · gpt-3.5-turbo
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    1334
    +0.07
    0.3%
    1034
    +0.07
    0.2%
    31
    +0.06
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    14
    +0.07
    0.02
    174
    +0.07
    0.02
    1681
    +0.06
    0.02
    Negative Logits
    <bos>
    -1.04
    /**
    -0.60
    WriteLiteral
    -0.58
    <?
    -0.56
    ⓧ
    -0.52
    //
    -0.52
    export
    -0.52
     relax
    -0.51
     jo
    -0.51
    join
    -0.51
    POSITIVE LOGITS
     tower
    2.68
     Tower
    2.49
     towers
    2.45
    Tower
    2.36
    tower
    2.27
     TOWER
    2.15
     Towers
    2.11
    towers
    2.00
     torre
    1.60
     torres
    1.37
    Activations Density 0.132%

    No Known Activations