© 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. 9500
    Prev
    Next
    INDEX
    Explanations

    words related to cultural practices and historical events, specifically around traditions like smoking meat and the origins of certain customs or names

    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
    496
    +0.12
    0.5%
    223
    +0.11
    0.5%
    103
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    103
    +0.12
    0.04
    507
    +0.11
    0.04
    609
    +0.10
    0.03
    Negative Logits
    htaking
    -0.85
    ertodd
    -0.78
    igslist
    -0.76
    apixel
    -0.74
     unloaded
    -0.73
    Recommend
    -0.71
    ggle
    -0.71
    ettings
    -0.70
    vidia
    -0.67
    ailable
    -0.67
    POSITIVE LOGITS
    ËĪ
    1.35
     slang
    1.22
    tymology
    1.20
     noun
    1.14
     pronunciation
    1.14
     plural
    1.11
     Meaning
    1.07
     suffix
    1.06
     Sanskrit
    1.05
     Latin
    1.05
    Activations Density 0.288%

    No Known Activations