© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 8901
    Prev
    Next
    INDEX
    Explanations

    proper nouns and technical terms related to current events and policies

    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
    430
    +0.15
    0.7%
    729
    +0.10
    0.5%
    723
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    430
    +0.15
    0.06
    306
    +0.10
    0.04
    136
    +0.10
    0.04
    Negative Logits
    itialized
    -0.66
    bler
    -0.66
    erate
    -0.65
    ership
    -0.63
    ggle
    -0.63
     gib
    -0.62
    gery
    -0.61
    Control
    -0.61
    ospel
    -0.60
     perpetual
    -0.60
    POSITIVE LOGITS
     afternoon
    0.90
    night
    0.85
     evening
    0.78
     announcing
    0.78
    days
    0.76
    flower
    0.76
     unveiling
    0.75
     morning
    0.71
    ghan
    0.71
     marked
    0.70
    Activations Density 0.073%

    No Known Activations