© 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. 15332
    Prev
    Next
    INDEX
    Explanations

    phrases with specific words related to a specific story or character

    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
    362
    +0.20
    0.9%
    393
    +0.15
    0.7%
    327
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.20
    0.11
    16
    +0.15
    0.06
    393
    +0.10
    0.06
    Negative Logits
     accordingly
    -0.82
    brate
    -0.79
    heit
    -0.75
    edly
    -0.75
    nesty
    -0.74
    seekers
    -0.74
    nance
    -0.73
    ulner
    -0.73
    paio
    -0.72
    LEASE
    -0.72
    POSITIVE LOGITS
     midst
    1.32
     guise
    1.08
     aftermath
    1.02
     vicinity
    0.96
     nutshell
    0.94
     afterlife
    0.94
     foreground
    0.93
     vein
    0.93
     Philippines
    0.92
     middle
    0.92
    Activations Density 0.247%

    No Known Activations