© 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. 16954
    Prev
    Next
    INDEX
    Explanations

    sections related to architectural features and detailed descriptions of specific locations

    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
    216
    +0.11
    0.5%
    752
    +0.10
    0.5%
    504
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    11
    +0.11
    0.01
    752
    +0.10
    0.01
    504
    +0.10
    0.01
    Negative Logits
     Annotations
    -0.69
    DIT
    -0.67
    EED
    -0.66
    lyak
    -0.66
    ossibility
    -0.66
    inational
    -0.66
    CHAT
    -0.66
    olitics
    -0.65
     Copy
    -0.63
    olean
    -0.61
    POSITIVE LOGITS
    ways
    1.48
    owed
    1.38
    marks
    1.26
    iday
    1.16
    ows
    1.15
    way
    1.06
    ibur
    1.02
     hall
    1.00
     halls
    0.96
    room
    0.94
    Activations Density 0.018%

    No Known Activations