© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 2165
    Prev
    Next
    INDEX
    Explanations

    descriptions of physical locations and situations

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    420
    +0.16
    0.7%
    281
    +0.11
    0.5%
    333
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.16
    0.07
    469
    +0.11
    0.03
    251
    +0.10
    0.04
    Negative Logits
    phabet
    -0.79
    Result
    -0.78
     enqu
    -0.78
     congr
    -0.72
    ewitness
    -0.68
     Nieto
    -0.67
    ="#
    -0.67
    intent
    -0.66
    ocide
    -0.64
    wow
    -0.64
    POSITIVE LOGITS
     rented
    1.19
     attic
    1.11
     tents
    1.07
     basement
    1.06
     cramped
    1.05
     dorm
    1.04
     confines
    1.03
     closet
    1.03
     padded
    0.99
     barracks
    0.99
    Activations Density 0.385%

    No Known Activations