© 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. 22168
    Prev
    Next
    INDEX
    Explanations

    references to personal pronouns indicating identity or presence

    oai_token-act-pair · gpt-4o-miniTriggered 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
    497
    +0.10
    0.5%
    126
    +0.10
    0.4%
    423
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    762
    +0.10
    0.04
    112
    +0.10
    0.03
    126
    +0.09
    0.02
    Negative Logits
    ģĸ
    -0.73
    ente
    -0.71
    ĨĴ
    -0.70
    ĵ
    -0.68
    ©¶æ¥µ
    -0.63
     Nights
    -0.62
    icion
    -0.62
    isma
    -0.60
    Panel
    -0.59
    creation
    -0.59
    POSITIVE LOGITS
     alive
    0.87
     smiling
    0.82
     unfold
    0.78
     perform
    0.77
    selves
    0.77
     interact
    0.76
     perched
    0.75
     silhou
    0.75
     grinning
    0.74
     slumped
    0.74
    Activations Density 0.109%

    No Known Activations