© 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. 5550
    Prev
    Next
    INDEX
    Explanations

    comparisons and contrasts using the phrase "like X, Y"

    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
    373
    +0.23
    1.1%
    464
    +0.12
    0.6%
    459
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    655
    +0.23
    0.06
    203
    +0.12
    0.04
    756
    +0.10
    -0.02
    Negative Logits
    Alert
    -0.65
    untled
    -0.62
    Continue
    -0.62
    ]).
    -0.61
    igslist
    -0.58
    Option
    -0.57
    UMP
    -0.56
    )).
    -0.54
    sbm
    -0.54
    Attempt
    -0.53
    POSITIVE LOGITS
     intrinsically
    0.84
     inherently
    0.83
     insofar
    0.83
     its
    0.77
     belongs
    0.74
     fundamentally
    0.72
     governs
    0.67
     finite
    0.66
     intrinsic
    0.64
     evolution
    0.64
    Activations Density 0.709%

    No Known Activations