© 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 49152
    9. 8-RES_FS49152-JB
    11. 45566
    Prev
    Next
    INDEX
    Explanations

    phrases expressing belief, hope, or concern about future outcomes

    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
    718
    +0.17
    0.8%
    608
    +0.14
    0.7%
    150
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    718
    +0.17
    0.02
    608
    +0.14
    0.02
    150
    +0.11
    0.02
    Negative Logits
    quickShipAvailable
    -0.74
     Appears
    -0.69
    ologies
    -0.68
    soever
    -0.67
    ONSORED
    -0.67
    thumbnails
    -0.66
     Laksh
    -0.65
    IDS
    -0.64
     Blitz
    -0.62
    yss
    -0.62
    POSITIVE LOGITS
     believe
    1.14
     doubt
    1.08
     distrust
    1.04
     mistrust
    1.02
     revisit
    1.00
     dislike
    0.97
     disbel
    0.96
     consider
    0.94
     celebrate
    0.94
     reconsider
    0.91
    Activations Density 0.030%

    No Known Activations