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 24576
    5. 8-RES_FS24576-JB
    6. 23728
    Prev
    Next
    INDEX
    Explanations

    references to existence and changes over time

    oai_token-act-pair · gpt-4o-miniTriggered 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
    481
    +0.14
    0.6%
    15
    +0.10
    0.4%
    276
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    455
    +0.14
    0.04
    341
    +0.10
    0.03
    15
    +0.09
    0.04
    Negative Logits
     Norn
    -0.68
     Lover
    -0.66
     Prices
    -0.63
     vict
    -0.63
     bidder
    -0.60
    éļ
    -0.60
    idad
    -0.59
     barr
    -0.59
     Fighters
    -0.58
     Tend
    -0.58
    POSITIVE LOGITS
     stemmed
    0.89
     revolves
    0.84
     stems
    0.83
     attributable
    0.76
     consists
    0.74
     consisted
    0.74
     relate
    0.74
    related
    0.71
     concentrated
    0.71
     relates
    0.69
    Activations Density 0.258%

    No Known Activations