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

    Neuronpedia

    Jacobian LensNEW
    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. 18150
    Prev
    Next
    INDEX
    Explanations

    names of individuals, likely indicating experts or researchers in a field related to health, environment, or policy

    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
    659
    +0.13
    0.6%
    373
    +0.10
    0.5%
    430
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.13
    0.08
    232
    +0.10
    0.06
    81
    +0.10
    0.06
    Negative Logits
    reddit
    -0.68
    pmwiki
    -0.66
    tumblr
    -0.66
     initiating
    -0.65
    DragonMagazine
    -0.64
     sabot
    -0.63
    Christmas
    -0.63
     apologizing
    -0.62
     revenge
    -0.62
     lockout
    -0.62
    POSITIVE LOGITS
     Shapiro
    1.03
     McInt
    1.00
     Kaplan
    0.99
     Greenberg
    0.98
     Feldman
    0.96
     Cohen
    0.95
     McK
    0.94
     Sachs
    0.93
     Becker
    0.93
     Levin
    0.92
    Activations Density 0.207%

    No Known Activations