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. 26433
    Prev
    Next
    INDEX
    Explanations

    technical terms related to computer systems and applications

    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
    308
    +0.12
    0.5%
    540
    +0.12
    0.5%
    423
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.12
    0.04
    547
    +0.12
    0.03
    39
    +0.10
    0.04
    Negative Logits
    pread
    -0.74
    ricular
    -0.69
    +=
    -0.67
    laus
    -0.65
    baugh
    -0.63
    Magn
    -0.62
    ometric
    -0.61
    omething
    -0.60
    oph
    -0.60
    tell
    -0.60
    POSITIVE LOGITS
     archive
    0.79
     itself
    0.79
     database
    0.77
     marketplace
    0.71
     archives
    0.70
     system
    0.69
    .
    0.67
    rique
    0.67
     platform
    0.67
    grounds
    0.66
    Activations Density 0.129%

    No Known Activations