© 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. 7177
    Prev
    Next
    INDEX
    Explanations

    instances where someone is contemplating or making a decision

    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
    192
    +0.10
    0.4%
    533
    +0.10
    0.4%
    8
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    533
    +0.10
    0.02
    512
    +0.10
    0.02
    192
    +0.09
    0.01
    Negative Logits
    verts
    -0.79
    vez
    -0.69
    akia
    -0.67
    hiba
    -0.65
    phrine
    -0.63
    vert
    -0.63
    eph
    -0.62
    ethy
    -0.62
    washer
    -0.61
    loads
    -0.60
    POSITIVE LOGITS
     nothing
    0.83
     TAMADRA
    0.76
    yip
    0.74
     they
    0.73
     admitting
    0.64
     possessed
    0.64
     THEY
    0.62
     menacing
    0.62
    acan
    0.62
     incapable
    0.61
    Activations Density 0.027%

    No Known Activations