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. 23428
    Prev
    Next
    INDEX
    Explanations

    references to specific graphics card models and their technical details

    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
    357
    +0.14
    0.6%
    538
    +0.12
    0.6%
    468
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.03
    326
    +0.12
    0.02
    357
    +0.12
    0.02
    Negative Logits
    urst
    -0.67
    olicy
    -0.63
    estinal
    -0.61
    clair
    -0.60
     kettle
    -0.60
    akra
    -0.60
    iso
    -0.60
    perty
    -0.60
     corros
    -0.60
    organic
    -0.59
    POSITIVE LOGITS
    eering
    0.85
    sburg
    0.82
    ãĥ¼ãĥĨãĤ£
    0.81
    ãĥ¼ãĤ¯
    0.75
    istas
    0.75
    DragonMagazine
    0.73
    ians
    0.73
    eer
    0.73
    icular
    0.72
    eers
    0.71
    Activations Density 0.047%

    No Known Activations