© 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. 22364
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate progress or improvement in various contexts

    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
    523
    +0.12
    0.6%
    373
    +0.11
    0.5%
    412
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    523
    +0.12
    0.03
    412
    +0.11
    0.02
    106
    +0.09
    0.02
    Negative Logits
    ousse
    -0.63
    fed
    -0.62
    dropping
    -0.59
    Widget
    -0.58
    ateurs
    -0.56
    item
    -0.56
    urities
    -0.55
    ãĤ¦ãĤ¹
    -0.54
     ende
    -0.54
    resent
    -0.54
    POSITIVE LOGITS
     route
    0.90
     furthe
    0.84
     downhill
    0.78
     bye
    0.72
     mile
    0.70
     lengths
    0.69
     distance
    0.69
    WARD
    0.65
     Marathon
    0.65
     routes
    0.64
    Activations Density 0.085%

    No Known Activations