© 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. 18916
    Prev
    Next
    INDEX
    Explanations

    adverbs indicating a time frame or continuity

    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
    430
    +0.11
    0.5%
    331
    +0.10
    0.5%
    460
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    430
    +0.11
    0.06
    253
    +0.10
    0.04
    55
    +0.10
    0.05
    Negative Logits
     quot
    -0.68
     root
    -0.61
     cess
    -0.59
     palette
    -0.58
    Pg
    -0.57
     immensely
    -0.56
     lim
    -0.55
     "$:/
    -0.55
     Wikimedia
    -0.54
    ords
    -0.54
    POSITIVE LOGITS
    that
    0.83
    rand
    0.75
     congratulations
    0.74
     how
    0.73
    whether
    0.73
     why
    0.70
    why
    0.68
    what
    0.68
     whether
    0.68
    ãĤ¤ãĥĪ
    0.67
    Activations Density 0.217%

    No Known Activations