© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 7959
    Prev
    Next
    INDEX
    Explanations

    words related to fitting a description or role, especially in a professional or physical context

    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
    546
    +0.10
    0.5%
    123
    +0.10
    0.5%
    540
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    123
    +0.10
    0.02
    614
    +0.10
    0.01
    469
    +0.10
    0.01
    Negative Logits
    ãĥ£
    -0.67
     authorize
    -0.65
     inciting
    -0.63
     wrest
    -0.63
     unable
    -0.62
     refusing
    -0.61
    icides
    -0.61
     downed
    -0.61
    oking
    -0.61
    iger
    -0.60
    POSITIVE LOGITS
     precon
    1.06
     snug
    0.94
     criteria
    0.92
     stereotype
    0.90
     mold
    0.87
     neatly
    0.87
    ahime
    0.85
     niche
    0.85
     stereotypes
    0.84
     mould
    0.83
    Activations Density 0.080%

    No Known Activations