© 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. 1565
    Prev
    Next
    INDEX
    Explanations

    phrases related to light-dark cycles and daily rhythms

    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
    548
    +0.10
    0.4%
    71
    +0.10
    0.4%
    341
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    289
    +0.10
    0.01
    678
    +0.10
    0.02
    578
    +0.09
    0.01
    Negative Logits
    aroo
    -0.79
    士
    -0.78
    legate
    -0.72
    >>\
    -0.68
    FactoryReloaded
    -0.67
    =-=-
    -0.66
    Ö¼
    -0.65
    çĭ
    -0.64
    sbm
    -0.63
    rency
    -0.62
    POSITIVE LOGITS
    elight
    1.12
     candles
    1.08
     lamps
    0.99
    raviolet
    0.95
     lantern
    0.93
     shone
    0.93
     shine
    0.93
     brighter
    0.92
     ultraviolet
    0.92
     candle
    0.90
    Activations Density 0.136%

    No Known Activations