© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 1583
    Prev
    Next
    INDEX
    Explanations

    mentions of people in specific roles or professions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    447
    +0.14
    0.6%
    506
    +0.13
    0.6%
    371
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    506
    +0.14
    0.08
    371
    +0.13
    0.08
    204
    +0.11
    0.07
    Negative Logits
    inki
    -0.70
    fw
    -0.65
    Request
    -0.64
     ga
    -0.64
     captcha
    -0.61
     [];
    -0.61
    imester
    -0.61
    ascript
    -0.61
    ãģį
    -0.60
     rehearsal
    -0.60
    POSITIVE LOGITS
     extraord
    1.39
     Richard
    1.10
     Shaun
    1.08
     Larry
    1.04
     Eugene
    1.04
     Erik
    1.04
     Sebast
    1.02
     Stephen
    1.02
     Gina
    1.01
     Colin
    1.00
    Activations Density 0.415%

    No Known Activations