© 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. 546
    Prev
    Next
    INDEX
    Explanations

    phrases related to instructions or guides

    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
    636
    +0.13
    0.6%
    362
    +0.11
    0.5%
    395
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    636
    +0.13
    0.07
    362
    +0.11
    0.07
    124
    +0.11
    0.05
    Negative Logits
    WARE
    -0.73
    step
    -0.72
    icion
    -0.69
    é¾įåĸļ士
    -0.69
    iculture
    -0.68
     Rooms
    -0.67
    yssey
    -0.67
    FIG
    -0.66
    FORMATION
    -0.65
    fg
    -0.65
    POSITIVE LOGITS
     pesky
    1.11
     altogether
    1.10
     unnecessary
    1.00
     inhib
    0.98
     superflu
    0.98
     unwanted
    0.96
     offending
    0.95
     barriers
    0.94
     entirety
    0.93
     incrim
    0.92
    Activations Density 0.353%

    No Known Activations