© 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 12288
    9. 8-RES_FS12288-JB
    11. 485
    Prev
    Next
    INDEX
    Explanations

    points in time or instances in a narrative

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    218
    +0.10
    0.5%
    651
    +0.10
    0.5%
    432
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    651
    +0.10
    0.04
    460
    +0.10
    0.03
    132
    +0.10
    0.03
    Negative Logits
    atts
    -0.65
    harm
    -0.63
     twins
    -0.61
    lez
    -0.60
    english
    -0.59
     twin
    -0.59
    mir
    -0.57
     subst
    -0.56
     Cities
    -0.55
     brothers
    -0.55
    POSITIVE LOGITS
    cture
    0.93
     onwards
    0.87
     onward
    0.84
    abouts
    0.76
    endment
    0.71
    SPONSORED
    0.68
    ozo
    0.67
    points
    0.64
    ebin
    0.61
    uncture
    0.61
    Activations Density 0.082%

    No Known Activations