© 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. 22594
    Prev
    Next
    INDEX
    Explanations

    occurrences of citations or references in a document

    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
    400
    +0.10
    0.5%
    364
    +0.10
    0.5%
    667
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    767
    +0.10
    0.02
    400
    +0.10
    0.02
    364
    +0.09
    0.02
    Negative Logits
    critical
    -0.60
    iencies
    -0.60
    Unity
    -0.58
    Dialogue
    -0.55
    igon
    -0.54
     Laurent
    -0.53
    ftime
    -0.52
    helle
    -0.51
    spot
    -0.50
    manac
    -0.50
    POSITIVE LOGITS
    )'
    1.51
    )(
    1.47
    )=
    1.38
    )-
    1.36
    )
    1.34
    )",
    1.30
    )"
    1.30
    )/
    1.29
    ):
    1.29
    ),"
    1.27
    Activations Density 0.095%

    No Known Activations