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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 24088
    Prev
    Next
    INDEX
    Explanations

    instances of the word "dip" and its variations, suggesting a focus on declines or reductions in various contexts

    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
    569
    +0.14
    0.6%
    35
    +0.14
    0.6%
    99
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    35
    +0.14
    0.01
    569
    +0.14
    0.01
    252
    +0.11
    0.01
    Negative Logits
    Reply
    -0.82
    Made
    -0.70
     ACTIONS
    -0.69
    fare
    -0.69
    eval
    -0.65
     WATCHED
    -0.65
    ########
    -0.64
     Crusade
    -0.63
    apes
    -0.63
    furt
    -0.63
    POSITIVE LOGITS
    shit
    1.15
     dipping
    0.95
    pty
    0.90
    irts
    0.87
    ole
    0.86
    stick
    0.85
    iphany
    0.85
    tera
    0.85
    rint
    0.82
    artisan
    0.81
    Activations Density 0.021%

    No Known Activations