© 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. 9535
    Prev
    Next
    INDEX
    Explanations

    the name of a specific person

    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
    659
    +0.11
    0.5%
    37
    +0.09
    0.4%
    238
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.11
    0.08
    232
    +0.09
    0.06
    284
    +0.09
    0.06
    Negative Logits
     Fargo
    -0.76
     underdog
    -0.70
     thr
    -0.69
     mainline
    -0.65
     stockpile
    -0.65
    IDENT
    -0.64
     grind
    -0.64
     brood
    -0.64
     roadside
    -0.63
    Vote
    -0.62
    POSITIVE LOGITS
    icz
    1.49
    akis
    1.35
    idis
    1.28
    anu
    1.26
    opoulos
    1.20
    oglu
    1.17
    enko
    1.15
    uez
    1.15
    owicz
    1.14
    cki
    1.13
    Activations Density 0.220%

    No Known Activations