© 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 49152
    9. 8-RES_FS49152-JB
    11. 25945
    Prev
    Next
    INDEX
    Explanations

    trends or themes related to "ch" or "chewing" in various contexts

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    174
    +0.13
    0.6%
    725
    +0.11
    0.5%
    526
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    174
    +0.13
    0.03
    725
    +0.11
    0.02
    243
    +0.11
    0.02
    Negative Logits
    PORT
    -0.78
    INGTON
    -0.78
    FORMATION
    -0.69
    ENCE
    -0.69
    ITED
    -0.69
    ragon
    -0.67
    OVA
    -0.67
    gaard
    -0.67
     Ajax
    -0.67
     Circus
    -0.65
    POSITIVE LOGITS
    amps
    1.25
    icago
    1.18
    inese
    1.18
    akra
    1.14
    owder
    1.14
    ipper
    1.13
    ipping
    1.12
    ieft
    1.11
    isel
    1.11
    asers
    1.11
    Activations Density 0.008%

    No Known Activations