© 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. 15354
    Prev
    Next
    INDEX
    Explanations

    phrases related to walking down specific streets

    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
    123
    +0.12
    0.5%
    459
    +0.11
    0.5%
    708
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    123
    +0.12
    0.03
    113
    +0.11
    0.02
    469
    +0.11
    0.02
    Negative Logits
    ²¾
    -0.77
    achev
    -0.76
    manship
    -0.76
    cffff
    -0.72
    catentry
    -0.72
    ingers
    -0.70
    bringing
    -0.70
    ilic
    -0.70
    pin
    -0.70
    fact
    -0.69
    POSITIVE LOGITS
     halls
    0.99
     corridors
    0.84
     doors
    0.84
     door
    0.79
     hallway
    0.79
     footprints
    0.79
     streets
    0.78
     hall
    0.77
     Ambro
    0.72
     gates
    0.72
    Activations Density 0.096%

    No Known Activations