© 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. 5932
    Prev
    Next
    INDEX
    Explanations

    negative sentiments or criticism towards a topic

    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
    203
    +0.09
    0.4%
    579
    +0.09
    0.4%
    307
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    610
    +0.09
    0.01
    663
    +0.09
    0.01
    509
    +0.09
    0.01
    Negative Logits
    lishing
    -0.67
    )=(
    -0.66
     Shades
    -0.61
     HIT
    -0.60
     Warning
    -0.59
    icut
    -0.57
     refresh
    -0.57
     ID
    -0.56
    STER
    -0.53
    APTER
    -0.53
    POSITIVE LOGITS
    be
    0.80
    soon
    0.75
    won
    0.75
    ignty
    0.71
    aughters
    0.71
    boa
    0.69
    cliffe
    0.69
    been
    0.67
    gered
    0.67
    minded
    0.67
    Activations Density 0.033%

    No Known Activations