© 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. 13122
    Prev
    Next
    INDEX
    Explanations

    contrasts and contradictions within a text

    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
    585
    +0.13
    0.6%
    244
    +0.11
    0.5%
    72
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    370
    +0.13
    0.02
    244
    +0.11
    0.03
    585
    +0.11
    0.03
    Negative Logits
     secondly
    -0.70
     later
    -0.62
    ":[
    -0.61
    cms
    -0.61
    borough
    -0.61
     afterwards
    -0.60
     Secondly
    -0.60
     afterward
    -0.60
     murd
    -0.58
    icter
    -0.58
    POSITIVE LOGITS
     only
    0.82
    Only
    0.73
    Recommend
    0.68
    oké
    0.67
    only
    0.66
    Supported
    0.64
    inoa
    0.64
    FTWARE
    0.62
    ipher
    0.61
    aido
    0.60
    Activations Density 0.238%

    No Known Activations