© 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. 9517
    Prev
    Next
    INDEX
    Explanations

    mentions of media coverage, website-related terms, strategy-related terms, interface-related terms, explanations, and announcements

    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
    481
    +0.18
    0.8%
    679
    +0.12
    0.6%
    705
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.18
    0.07
    679
    +0.12
    0.07
    540
    +0.08
    0.06
    Negative Logits
    nee
    -0.72
    itton
    -0.71
    eln
    -0.64
    ldon
    -0.59
    ãĥ©ãĥ³
    -0.58
    ometimes
    -0.57
    visors
    -0.56
    ipel
    -0.54
    asty
    -0.54
     beware
    -0.54
    POSITIVE LOGITS
     nor
    1.35
     whatsoever
    1.32
     anymore
    1.15
     anywhere
    0.87
     yet
    0.84
     either
    0.82
     satisfactory
    0.81
     forthcoming
    0.80
     besides
    0.76
    nor
    0.72
    Activations Density 0.392%

    No Known Activations