© 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 12288
    9. 8-RES_FS12288-JB
    11. 511
    Prev
    Next
    INDEX
    Explanations

    instances of spatial relationships, in particular, referring to the back of something

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    393
    +0.12
    0.6%
    373
    +0.11
    0.5%
    683
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.12
    0.04
    587
    +0.11
    0.04
    476
    +0.11
    0.03
    Negative Logits
     Cosponsors
    -0.79
    terday
    -0.74
    dayName
    -0.72
    BIL
    -0.70
    Quantity
    -0.69
    NL
    -0.67
    864
    -0.67
    20439
    -0.64
    Ô
    -0.62
    forts
    -0.61
    POSITIVE LOGITS
     each
    0.73
     scrimmage
    0.67
     the
    0.64
     Hawkins
    0.61
     Cour
    0.60
     thigh
    0.60
     abdomen
    0.58
     stairs
    0.58
     Alto
    0.58
     Gors
    0.57
    Activations Density 0.190%

    No Known Activations