Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 12081
    Prev
    Next
    INDEX
    Explanations

    concepts related to expectations and their impact on individuals and society

    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
    447
    +0.13
    0.6%
    762
    +0.10
    0.5%
    268
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    714
    +0.13
    0.02
    762
    +0.10
    0.04
    331
    +0.09
    0.03
    Negative Logits
    ongh
    -0.64
    ovember
    -0.60
    va
    -0.58
    pm
    -0.57
     Mayhem
    -0.57
    urations
    -0.56
    greSQL
    -0.56
    edin
    -0.55
    td
    -0.55
    racuse
    -0.55
    POSITIVE LOGITS
     nonetheless
    0.83
    !--
    0.73
    arently
    0.72
     preferably
    0.70
     whose
    0.67
     which
    0.66
    _-
    0.65
     incidentally
    0.65
    which
    0.64
    imately
    0.63
    Activations Density 0.197%

    No Known Activations