© 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. 24404
    Prev
    Next
    INDEX
    Explanations

    words related to assistance or support

    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
    327
    +0.10
    0.4%
    718
    +0.10
    0.4%
    504
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    327
    +0.10
    0.04
    112
    +0.10
    0.04
    504
    +0.09
    0.03
    Negative Logits
    iu
    -0.69
     Leaves
    -0.63
    é¾
    -0.61
    meat
    -0.61
     squared
    -0.58
    ata
    -0.58
    ircraft
    -0.58
     Mellon
    -0.57
    been
    -0.57
     Charl
    -0.57
    POSITIVE LOGITS
     alleviate
    1.09
     facilitate
    1.05
     stabilize
    0.99
     improve
    0.96
     propel
    0.96
     mitigate
    0.95
     illustrate
    0.94
     elevate
    0.93
     define
    0.93
     relieve
    0.93
    Activations Density 0.043%

    No Known Activations