© 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. 16517
    Prev
    Next
    INDEX
    Explanations

    numbers throughout the text

    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
    746
    +0.14
    0.7%
    598
    +0.12
    0.5%
    29
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    598
    +0.14
    0.04
    316
    +0.12
    0.03
    326
    +0.10
    0.03
    Negative Logits
    ications
    -0.78
    istical
    -0.73
     bleach
    -0.73
    ishes
    -0.73
    iated
    -0.68
    nces
    -0.68
    ition
    -0.68
    ioned
    -0.68
    ist
    -0.67
    iations
    -0.66
    POSITIVE LOGITS
    09
    1.12
    08
    1.09
    07
    1.06
    06
    1.04
    05
    1.02
    04
    1.01
    090
    0.99
    059
    0.99
    03
    0.98
    089
    0.98
    Activations Density 0.099%

    No Known Activations