© 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 EvalsExportsinterp-engineNEWAPI 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. 22434
    Prev
    Next
    INDEX
    Explanations

    instances of personal opinion phrasing in discussions

    oai_token-act-pair · gpt-4o-miniTriggered 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
    481
    +0.17
    0.8%
    104
    +0.10
    0.4%
    687
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    498
    +0.17
    0.03
    599
    +0.10
    0.02
    452
    +0.09
    0.02
    Negative Logits
    olor
    -0.71
     Others
    -0.69
     Explan
    -0.67
    Alert
    -0.65
    arine
    -0.65
    -+-+
    -0.65
    asionally
    -0.62
    advertising
    -0.61
    lees
    -0.60
    ships
    -0.60
    POSITIVE LOGITS
     instead
    0.74
    phasis
    0.72
    instead
    0.72
     emphasis
    0.68
    llah
    0.67
     opted
    0.65
     understatement
    0.65
     lucky
    0.64
    quickShipAvailable
    0.63
    iller
    0.63
    Activations Density 0.202%

    No Known Activations