© 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. 5110
    Prev
    Next
    INDEX
    Explanations

    combinations of words and phrases indicating a debate or argument

    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
    481
    +0.37
    1.8%
    314
    +0.15
    0.7%
    195
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    314
    +0.37
    -0.01
    738
    +0.15
    0.03
    195
    +0.10
    0.01
    Negative Logits
     constitu
    -0.70
     Azerb
    -0.67
     unres
    -0.60
     MSI
    -0.60
     Emirates
    -0.59
     NTS
    -0.58
     cumbers
    -0.58
     religiously
    -0.57
     Emir
    -0.57
     immediate
    -0.57
    POSITIVE LOGITS
    wolves
    0.87
    chin
    0.85
    wolf
    0.80
    lip
    0.80
    bt
    0.79
    omorph
    0.79
    vec
    0.79
    veh
    0.77
    wer
    0.77
    xit
    0.76
    Activations Density 0.866%

    No Known Activations