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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 19821
    Prev
    Next
    INDEX
    Explanations

    references to the meaning or significance of words or phrases

    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
    642
    +0.17
    0.8%
    712
    +0.11
    0.5%
    178
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    642
    +0.17
    0.02
    712
    +0.11
    0.01
    688
    +0.10
    0.01
    Negative Logits
    aliation
    -0.73
    asting
    -0.69
    emetery
    -0.67
    asted
    -0.65
    Tur
    -0.63
     Breed
    -0.63
    culosis
    -0.62
     Turk
    -0.61
    hani
    -0.58
    Monitor
    -0.57
    POSITIVE LOGITS
    fully
    1.53
    ful
    1.26
    fulness
    1.25
    lessness
    1.23
    FUL
    0.98
    lessly
    0.96
    sworth
    0.90
    bringer
    0.84
    full
    0.82
    terday
    0.82
    Activations Density 0.026%

    No Known Activations