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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 4206
    Prev
    Next
    INDEX
    Explanations

    social and cultural references with a focus on food and community interaction

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    659
    +0.16
    0.8%
    742
    +0.13
    0.6%
    493
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.16
    0.05
    742
    +0.13
    0.04
    493
    +0.09
    0.03
    Negative Logits
     boundaries
    -0.68
     Subtle
    -0.67
     prohibition
    -0.66
    éĹĺ
    -0.63
    ãĥ¼ãĥĨãĤ£
    -0.63
     compensated
    -0.62
     Prohibition
    -0.61
     correctness
    -0.61
    fulness
    -0.59
     absor
    -0.59
    POSITIVE LOGITS
    yrics
    1.28
    ONDON
    1.23
    ateral
    1.21
    aughter
    1.21
    OWER
    1.20
    incoln
    1.18
    ifestyle
    1.17
    UX
    1.13
    aptop
    1.12
    ocker
    1.11
    Activations Density 0.043%

    No Known Activations