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. 26315
    Prev
    Next
    INDEX
    Explanations

    references to valves and the company Valve

    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
    500
    +0.13
    0.6%
    288
    +0.12
    0.5%
    188
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    500
    +0.13
    0.01
    146
    +0.12
    0.01
    543
    +0.10
    0.01
    Negative Logits
    agy
    -0.82
    mares
    -0.78
    elf
    -0.74
    ocene
    -0.72
    ciating
    -0.70
    ointed
    -0.70
    edly
    -0.69
    escription
    -0.67
    iary
    -0.65
    alos
    -0.63
    POSITIVE LOGITS
     Studios
    0.85
     Machines
    0.77
     Corporation
    0.77
    Maps
    0.75
    Ins
    0.73
     Springs
    0.72
     Maps
    0.71
     Entertainment
    0.70
    Games
    0.70
     Presents
    0.69
    Activations Density 0.005%

    No Known Activations