© 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. 24484
    Prev
    Next
    INDEX
    Explanations

    the phrase "to" along with variations of its use in context, indicating obligations or actions

    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
    718
    +0.12
    0.5%
    620
    +0.10
    0.5%
    49
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    49
    +0.12
    0.04
    551
    +0.10
    0.04
    718
    +0.10
    0.03
    Negative Logits
     Cyp
    -0.74
    Motion
    -0.73
    interstitial
    -0.71
    Eva
    -0.71
     Crate
    -0.62
     Lif
    -0.61
    wikipedia
    -0.60
    handled
    -0.60
    ufact
    -0.60
     Cecil
    -0.59
    POSITIVE LOGITS
     worry
    1.43
     bother
    1.15
     rely
    1.08
     wait
    1.06
     anymore
    0.95
     endure
    0.94
     contend
    0.94
     compete
    0.94
     confront
    0.94
     intervene
    0.93
    Activations Density 0.063%

    No Known Activations