© 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 12288
    9. 8-RES_FS12288-JB
    11. 9374
    Prev
    Next
    INDEX
    Explanations

    requests for feedback or information from an audience

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    517
    +0.10
    0.4%
    726
    +0.10
    0.4%
    748
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    726
    +0.10
    0.03
    424
    +0.10
    0.02
    748
    +0.08
    0.02
    Negative Logits
    erva
    -0.64
    allowed
    -0.63
    ashi
    -0.62
    ©¶æ
    -0.61
    Rap
    -0.61
    ãĤ§
    -0.60
    ortality
    -0.60
    Dash
    -0.60
    olving
    -0.60
    rel
    -0.59
    POSITIVE LOGITS
     ASAP
    0.84
     beforehand
    0.79
     anytime
    0.77
     feedback
    0.76
     how
    0.76
     info
    0.73
     tid
    0.71
     compliments
    0.70
     suggestions
    0.70
     hello
    0.69
    Activations Density 0.079%

    No Known Activations