© 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. 5288
    Prev
    Next
    INDEX
    Explanations

    companies or organizations and their 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
    469
    +0.13
    0.6%
    261
    +0.12
    0.5%
    362
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    469
    +0.13
    0.04
    123
    +0.12
    0.05
    261
    +0.10
    0.03
    Negative Logits
    umatic
    -0.88
     hurled
    -0.85
    xual
    -0.80
    phabet
    -0.80
     wielded
    -0.74
     handed
    -0.72
     consumed
    -0.71
    aped
    -0.71
    handled
    -0.69
     recip
    -0.69
    POSITIVE LOGITS
     downtown
    1.12
     Alexandria
    1.11
     Cincinnati
    1.10
     Arlington
    1.09
     Pasadena
    1.08
     Omaha
    1.08
     Pittsburgh
    1.08
     Cambridge
    1.07
     Bangalore
    1.06
     Vienna
    1.05
    Activations Density 0.132%

    No Known Activations