© 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. 22803
    Prev
    Next
    INDEX
    Explanations

    future tense statements indicating intentions or plans

    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
    498
    +0.10
    0.4%
    455
    +0.09
    0.4%
    665
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.10
    0.05
    49
    +0.09
    0.04
    354
    +0.09
    0.03
    Negative Logits
    Alert
    -0.74
     Kills
    -0.69
     intolerable
    -0.69
     outp
    -0.67
     Typhoon
    -0.66
     Allies
    -0.62
     Canaveral
    -0.62
     pandemonium
    -0.61
     Crimson
    -0.61
    Jar
    -0.61
    POSITIVE LOGITS
     summarize
    0.99
     assume
    0.97
     introduce
    0.96
     briefly
    0.94
     discuss
    0.92
     explore
    0.92
     analyse
    0.91
     illustrate
    0.91
     summar
    0.90
     delve
    0.89
    Activations Density 0.118%

    No Known Activations