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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 18836
    Prev
    Next
    INDEX
    Explanations

    information or informing activities

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    679
    +0.14
    0.7%
    756
    +0.10
    0.5%
    721
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    59
    +0.14
    0.02
    679
    +0.10
    0.02
    417
    +0.09
    0.02
    Negative Logits
    ingu
    -0.69
    enf
    -0.67
     vain
    -0.64
    mad
    -0.62
    neys
    -0.58
    amiya
    -0.58
    nam
    -0.58
    opes
    -0.57
    dogs
    -0.56
    urdue
    -0.56
    POSITIVE LOGITS
    ally
    1.09
     consent
    0.89
    omic
    0.87
    isance
    0.84
    tale
    0.83
    ingly
    0.82
     Consent
    0.80
     inform
    0.76
     constitu
    0.76
    umin
    0.76
    Activations Density 0.033%

    No Known Activations