© 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. 13074
    Prev
    Next
    INDEX
    Explanations

    pros and cons in texts along with positive aspects and advantages

    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
    369
    +0.12
    0.5%
    745
    +0.10
    0.5%
    552
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    444
    +0.12
    0.03
    285
    +0.10
    0.03
    724
    +0.10
    0.03
    Negative Logits
    uled
    -0.69
    dit
    -0.64
    bage
    -0.63
     chars
    -0.61
    lio
    -0.60
     boarding
    -0.59
    bus
    -0.59
     Donation
    -0.59
    inarily
    -0.58
     Payment
    -0.57
    POSITIVE LOGITS
     outweigh
    1.25
     outwe
    1.19
     conferred
    0.95
     Flavoring
    0.86
     inherent
    0.79
    itive
    0.78
    iveness
    0.74
     bestowed
    0.74
    ensical
    0.74
     afforded
    0.72
    Activations Density 0.098%

    No Known Activations