© 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. 37170
    Prev
    Next
    INDEX
    Explanations

    phrases that reference knowledge or information

    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
    638
    +0.11
    0.5%
    161
    +0.10
    0.5%
    709
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    638
    +0.11
    0.03
    155
    +0.10
    0.02
    161
    +0.09
    0.02
    Negative Logits
     teasp
    -0.99
    amiya
    -0.75
    clave
    -0.74
    ©¶æ¥µ
    -0.74
    aunder
    -0.73
    OGR
    -0.72
    cknow
    -0.71
    à¥
    -0.71
    provided
    -0.71
    cue
    -0.70
    POSITIVE LOGITS
     how
    0.98
     ourselves
    0.93
     them
    0.82
     him
    0.77
     these
    0.76
     what
    0.74
     Himself
    0.74
     our
    0.73
     regards
    0.73
     why
    0.72
    Activations Density 0.049%

    No Known Activations