© 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 49152
    9. 8-RES_FS49152-JB
    11. 41732
    Prev
    Next
    INDEX
    Explanations

    web-related addresses or links

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    373
    +0.15
    0.7%
    314
    +0.11
    0.5%
    57
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    334
    +0.15
    0.01
    687
    +0.11
    0.01
    57
    +0.09
    0.01
    Negative Logits
     proportion
    -0.69
     expulsion
    -0.67
     Loans
    -0.60
     Labrador
    -0.60
     contrary
    -0.59
     quart
    -0.58
     BART
    -0.58
     IO
    -0.58
     Bend
    -0.57
     Ballard
    -0.57
    POSITIVE LOGITS
    www
    1.38
    natureconservancy
    1.13
    ://
    1.03
    forums
    1.03
    forum
    1.02
    ww
    0.96
    youtu
    0.95
     www
    0.93
    ibaba
    0.93
    youtube
    0.91
    Activations Density 0.006%

    No Known Activations