© 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. 16221
    Prev
    Next
    INDEX
    Explanations

    text related to discussing controversial topics online and potential legal issues surrounding publishing

    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
    481
    +0.22
    1.0%
    621
    +0.10
    0.5%
    568
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    384
    +0.22
    0.03
    393
    +0.10
    0.05
    645
    +0.10
    0.03
    Negative Logits
    ovember
    -0.79
    akra
    -0.71
     citation
    -0.66
    estone
    -0.63
    inez
    -0.60
    cu
    -0.60
    Salt
    -0.59
     laun
    -0.58
    buster
    -0.57
    enne
    -0.57
    POSITIVE LOGITS
     ones
    1.28
    those
    1.21
     those
    1.11
     losers
    1.03
     newcomers
    0.94
     individuals
    0.91
     others
    0.87
     Those
    0.86
     folks
    0.86
    Those
    0.82
    Activations Density 0.349%

    No Known Activations