© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 3350
    Prev
    Next
    INDEX
    Explanations

    phrases related to law enforcement activities, specifically mentioning patrol and cruiser

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    481
    +0.12
    0.6%
    447
    +0.12
    0.5%
    113
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    104
    +0.12
    0.01
    575
    +0.12
    0.01
    288
    +0.11
    0.01
    Negative Logits
    xual
    -0.68
    hur
    -0.67
    igmat
    -0.67
     mortg
    -0.66
     nerv
    -0.65
     Centauri
    -0.63
    Torrent
    -0.61
    rencies
    -0.61
     toast
    -0.60
    CRIP
    -0.60
    POSITIVE LOGITS
     patrol
    1.12
     patrols
    1.05
    men
    0.94
     patrolling
    0.89
     duty
    0.88
    man
    0.88
     Patrol
    0.87
     officers
    0.83
     officer
    0.79
     checkpoints
    0.78
    Activations Density 0.026%

    No Known Activations