© 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. 24118
    Prev
    Next
    INDEX
    Explanations

    verbs that indicate creation, impact, or ethical considerations 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
    447
    +0.13
    0.6%
    252
    +0.10
    0.4%
    758
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    761
    +0.13
    0.05
    642
    +0.10
    0.04
    665
    +0.09
    0.05
    Negative Logits
     Reported
    -0.78
     Known
    -0.70
    bats
    -0.68
    reported
    -0.63
    christ
    -0.62
    picked
    -0.62
     Browse
    -0.62
     Seeking
    -0.62
    approved
    -0.61
    inian
    -0.61
    POSITIVE LOGITS
     tremend
    0.91
     itself
    0.81
     misunderstand
    0.81
     alot
    0.78
     valuable
    0.76
     imagination
    0.74
     morale
    0.74
     tremendous
    0.72
     empathy
    0.72
     enormously
    0.72
    Activations Density 0.334%

    No Known Activations