© 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. 4874
    Prev
    Next
    INDEX
    Explanations

    phrases related to evaluation or assessment of something

    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
    55
    +0.14
    0.7%
    102
    +0.14
    0.7%
    481
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.14
    0.07
    55
    +0.14
    0.07
    566
    +0.13
    0.05
    Negative Logits
    udeau
    -0.71
    isions
    -0.69
    ainers
    -0.66
     remarks
    -0.66
    itches
    -0.65
     Carib
    -0.64
    ffe
    -0.63
     Bucc
    -0.63
     Colomb
    -0.63
     Guys
    -0.62
    POSITIVE LOGITS
     reminiscent
    1.02
     supposed
    0.98
     designed
    0.96
     antit
    0.94
     indistinguishable
    0.93
     compatible
    0.92
     meant
    0.91
    cedented
    0.90
     destined
    0.90
     neither
    0.90
    Activations Density 0.177%

    No Known Activations