© 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. 22472
    Prev
    Next
    INDEX
    Explanations

    instances of the word "left" across 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
    289
    +0.09
    0.4%
    624
    +0.09
    0.4%
    50
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    468
    +0.09
    0.02
    50
    +0.09
    0.02
    650
    +0.09
    0.02
    Negative Logits
    wcsstore
    -0.65
    issance
    -0.63
     GENERAL
    -0.62
    andise
    -0.62
     Offense
    -0.61
    antz
    -0.60
    Temperature
    -0.60
     mesh
    -0.58
    NT
    -0.58
    insula
    -0.58
    POSITIVE LOGITS
    overs
    1.26
     untreated
    1.04
     undone
    1.03
     behind
    0.96
    wing
    0.90
     unanswered
    0.86
     intact
    0.85
     untouched
    0.85
     unfinished
    0.85
    behind
    0.85
    Activations Density 0.029%

    No Known Activations