© 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. 1643
    Prev
    Next
    INDEX
    Explanations

    the word "Utilization" in various forms and 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
    23
    +0.13
    0.6%
    311
    +0.13
    0.6%
    420
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    311
    +0.13
    0.01
    23
    +0.13
    0.02
    232
    +0.12
    0.01
    Negative Logits
     Carnival
    -0.71
    pmwiki
    -0.67
    boards
    -0.65
    manship
    -0.61
     Cosponsors
    -0.60
     thinner
    -0.59
    wcs
    -0.57
     trademarks
    -0.57
     Sins
    -0.57
     CPC
    -0.57
    POSITIVE LOGITS
    ilitarian
    1.37
    rue
    1.16
    ilities
    1.10
    ility
    1.10
    opia
    1.07
    tering
    1.05
    terness
    1.04
    rient
    1.01
    rients
    1.01
    hered
    0.99
    Activations Density 0.023%

    No Known Activations