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. 23845
    Prev
    Next
    INDEX
    Explanations

    statements regarding the limitations of scientific methods to explain human behavior and social phenomena

    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
    87
    +0.22
    1.1%
    756
    +0.16
    0.8%
    762
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    756
    +0.22
    -0.01
    87
    +0.16
    -0.03
    762
    +0.13
    0.07
    Negative Logits
     Hogan
    -0.80
    ometown
    -0.77
    Metro
    -0.77
    laun
    -0.77
     CTR
    -0.76
     Boom
    -0.71
    Rocket
    -0.70
     marquee
    -0.69
    CDC
    -0.69
    izzle
    -0.69
    POSITIVE LOGITS
     epist
    1.12
     philosophers
    1.07
     spiritual
    1.05
     Hegel
    1.03
     consciousness
    1.03
     divine
    1.01
     presupp
    1.00
     unconsciously
    0.99
     emanc
    0.98
     obedience
    0.98
    Activations Density 0.779%

    No Known Activations