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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 19872
    Prev
    Next
    INDEX
    Explanations

    mentions of various substances and their effects on individuals

    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
    447
    +0.31
    1.5%
    481
    +0.15
    0.7%
    252
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    461
    +0.31
    0.06
    288
    +0.15
    0.07
    592
    +0.11
    0.05
    Negative Logits
    ointed
    -0.73
    20439
    -0.71
     Plymouth
    -0.67
    olulu
    -0.66
    uminati
    -0.64
     Ern
    -0.59
    ¶ħ
    -0.59
     UNC
    -0.58
     NRL
    -0.57
    mitt
    -0.57
    POSITIVE LOGITS
     exists
    1.09
     represents
    1.02
     behaves
    1.00
     itself
    0.99
     isn
    0.98
     existed
    0.95
     tends
    0.93
     is
    0.92
     occupies
    0.89
     sucks
    0.89
    Activations Density 0.333%

    No Known Activations