© 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. 2655
    Prev
    Next
    INDEX
    Explanations

    services or support-related words, such as "legal services," "protection," "medical care," and "advice."

    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
    189
    +0.14
    0.6%
    439
    +0.13
    0.6%
    447
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    439
    +0.14
    0.05
    225
    +0.13
    0.05
    575
    +0.13
    0.06
    Negative Logits
    aten
    -0.66
    Flickr
    -0.56
     Seas
    -0.56
     Wikimedia
    -0.56
     senses
    -0.56
    assing
    -0.56
     overe
    -0.55
     discarded
    -0.54
    asses
    -0.54
     nerves
    -0.54
    POSITIVE LOGITS
     unto
    0.89
     to
    0.80
     onto
    0.78
    services
    0.72
    vice
    0.71
    enance
    0.70
    PsyNetMessage
    0.69
     whilst
    0.69
     towards
    0.68
     alongside
    0.68
    Activations Density 0.310%

    No Known Activations