© 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. 37266
    Prev
    Next
    INDEX
    Explanations

    terms related to migration and related processes

    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
    288
    +0.17
    0.8%
    447
    +0.16
    0.8%
    575
    +0.16
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.17
    0.02
    619
    +0.16
    0.02
    288
    +0.16
    0.02
    Negative Logits
    amina
    -0.76
    atur
    -0.70
    Cola
    -0.70
    rah
    -0.69
    reme
    -0.68
    terness
    -0.68
    NESS
    -0.66
    stood
    -0.63
    ihil
    -0.63
    ubes
    -0.63
    POSITIVE LOGITS
     flows
    0.78
     destinations
    0.74
     patterns
    0.72
     routes
    0.72
     migrating
    0.68
     movements
    0.67
     migr
    0.67
     paths
    0.67
     hubs
    0.64
     away
    0.64
    Activations Density 0.012%

    No Known Activations