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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 42862
    Prev
    Next
    INDEX
    Explanations

    references to collective actions or efforts

    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
    +0.37
    1.9%
    480
    +0.29
    1.5%
    481
    +0.23
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    266
    +0.37
    0.04
    480
    +0.29
    0.01
    421
    +0.23
    0.04
    Negative Logits
     withd
    -0.65
     redes
    -0.63
    æ©
    -0.58
    é¾įå
    -0.55
    essage
    -0.55
    ļéĨĴ
    -0.54
     Cooke
    -0.53
     Âł Âł Âł Âł Âł Âł Âł Âł
    -0.52
    req
    -0.51
    ription
    -0.51
    POSITIVE LOGITS
    "—
    0.91
    "?
    0.90
    "]
    0.88
    ,'"
    0.88
    %"
    0.87
    ")
    0.84
    zbollah
    0.80
    ,"
    0.79
    "),
    0.78
    .")
    0.78
    Activations Density 0.214%

    No Known Activations