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. 26369
    Prev
    Next
    INDEX
    Explanations

    references to the White House

    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
    167
    +0.11
    0.5%
    574
    +0.10
    0.5%
    57
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    574
    +0.11
    0.03
    57
    +0.10
    0.03
    175
    +0.10
    0.02
    Negative Logits
    Downloadha
    -0.83
    anwhile
    -0.82
    yrinth
    -0.78
    aterasu
    -0.78
    awaru
    -0.74
    ngth
    -0.72
    rawdownloadcloneembedreportprint
    -0.72
    ITAL
    -0.70
    REM
    -0.69
    ript
    -0.66
    POSITIVE LOGITS
     House
    1.15
    house
    1.13
     Sox
    1.09
    hall
    1.05
    ewater
    0.97
     supremacist
    0.93
    caps
    0.92
    House
    0.91
    hurst
    0.91
     supremacists
    0.86
    Activations Density 0.016%

    No Known Activations