© 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. 2944
    Prev
    Next
    INDEX
    Explanations

    phrases related to belonging and community involvement

    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
    393
    +0.11
    0.5%
    481
    +0.11
    0.5%
    362
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    210
    +0.11
    0.02
    362
    +0.11
    0.03
    695
    +0.10
    0.02
    Negative Logits
    oir
    -0.82
    iste
    -0.72
     Appears
    -0.69
     lied
    -0.68
    ocry
    -0.66
    ynes
    -0.65
    lah
    -0.64
     stars
    -0.63
    acs
    -0.62
    nce
    -0.61
    POSITIVE LOGITS
     ongoing
    0.94
     preparations
    0.90
     Operation
    0.84
    interstitial
    0.83
    Operation
    0.81
    lining
    0.73
     precaution
    0.73
     negotiations
    0.72
     their
    0.72
     broader
    0.71
    Activations Density 0.069%

    No Known Activations