© 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. 19792
    Prev
    Next
    INDEX
    Explanations

    verbs related to actions or decisions being performed

    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
    665
    +0.12
    0.6%
    95
    +0.11
    0.5%
    177
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.12
    0.09
    611
    +0.11
    0.06
    70
    +0.10
    0.06
    Negative Logits
    è¦ļéĨĴ
    -0.76
    emonic
    -0.67
    uras
    -0.65
    ascus
    -0.65
    >>\
    -0.63
    cosystem
    -0.63
    urus
    -0.61
     supposed
    -0.61
    osure
    -0.60
    Cas
    -0.60
    POSITIVE LOGITS
     tomorrow
    0.87
     whoever
    0.76
    fully
    0.76
     hereafter
    0.75
     thee
    0.75
     whatever
    0.73
    enance
    0.72
     someday
    0.70
     whichever
    0.70
     forever
    0.69
    Activations Density 0.446%

    No Known Activations