© 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 49152
    9. 8-RES_FS49152-JB
    11. 37294
    Prev
    Next
    INDEX
    Explanations

    identifiers or references to specific content

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    373
    +0.12
    0.5%
    300
    +0.12
    0.5%
    515
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.12
    0.02
    640
    +0.12
    0.01
    573
    +0.11
    0.01
    Negative Logits
     combust
    -0.80
    senal
    -0.79
     closet
    -0.74
     domestically
    -0.72
     lull
    -0.69
     paycheck
    -0.67
     corrid
    -0.67
    intendent
    -0.65
     exha
    -0.65
     overseas
    -0.65
    POSITIVE LOGITS
    UTC
    1.29
    ð
    0.78
    ajor
    0.77
    Hello
    0.77
    Hi
    0.76
     Firstly
    0.76
     Explain
    0.75
    itars
    0.75
     âĨij
    0.73
     Presumably
    0.72
    Activations Density 0.013%

    No Known Activations