© 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. 21023
    Prev
    Next
    INDEX
    Explanations

    phrases related to commands or requests

    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
    665
    +0.21
    1.0%
    217
    +0.13
    0.6%
    718
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.21
    0.07
    718
    +0.13
    0.04
    608
    +0.11
    0.04
    Negative Logits
    ileaks
    -0.61
    nown
    -0.61
    hur
    -0.60
     genesis
    -0.59
     unintentional
    -0.58
    tun
    -0.58
    fortunately
    -0.58
     ed
    -0.57
    raft
    -0.57
     seq
    -0.56
    POSITIVE LOGITS
     participate
    1.15
     perform
    1.13
     undertake
    1.09
     submit
    1.06
     recite
    1.04
     testify
    1.03
     uphold
    1.02
     undergo
    1.02
     partake
    1.01
     attend
    1.01
    Activations Density 0.111%

    No Known Activations