© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Initial Gemma-2B-IT SAEs
    5. Gemma-2B-IT
    7. Residual Stream
    9. 12-RES-JB
    11. 4412
    Prev
    Next
    INDEX
    Explanations

    texts discussing food and meals

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Hook Layer
    12
    Architecture
    standard
    Context Size
    1,024
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    1013
    +0.17
    0.5%
    674
    +0.09
    0.3%
    936
    +0.08
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1013
    +0.17
    0.06
    1336
    +0.09
    0.05
    1166
    +0.08
    0.06
    Negative Logits
     ?...
    -1.31
     unwarran
    -1.28
     increa
    -1.28
     Cfr
    -1.27
     encomp
    -1.26
     emphat
    -1.25
     Ibidem
    -1.25
     disagre
    -1.25
     §.
    -1.24
     affor
    -1.23
    POSITIVE LOGITS
    <bos>
    0.94
     morning
    0.86
     breakfast
    0.81
     daily
    0.73
     mornings
    0.71
    morning
    0.67
     lunch
    0.64
     afternoon
    0.60
    breakfast
    0.58
     usually
    0.57
    Activations Density 0.433%

    No Known Activations