© 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 EvalsExportsAPI 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. 6387
    Prev
    Next
    INDEX
    Explanations

    invitations or encouragements

    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
    373
    +0.16
    0.7%
    590
    +0.09
    0.4%
    279
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    269
    +0.16
    0.02
    654
    +0.09
    0.02
    168
    +0.09
    0.02
    Negative Logits
    ãĤ©
    -0.66
    nesota
    -0.65
    ython
    -0.63
    acters
    -0.63
    cpp
    -0.59
    ACTED
    -0.59
    ĻĤ
    -0.58
    itans
    -0.57
    ults
    -0.57
    utton
    -0.56
    POSITIVE LOGITS
     behalf
    1.07
    erous
    1.04
     board
    0.98
    shore
    0.95
    eness
    0.94
    stage
    0.91
     stage
    0.87
    etime
    0.86
    screen
    0.85
    sets
    0.83
    Activations Density 0.050%

    No Known Activations