© 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 24576
    9. 8-RES_FS24576-JB
    11. 7146
    Prev
    Next
    INDEX
    Explanations

    phrases related to technological processes and instructions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    565
    +0.10
    0.5%
    534
    +0.10
    0.5%
    560
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    300
    +0.10
    0.04
    545
    +0.10
    0.03
    566
    +0.10
    0.04
    Negative Logits
     surpr
    -0.63
     alternatively
    -0.62
     envy
    -0.61
     differently
    -0.61
    always
    -0.61
    ILY
    -0.60
    ItemImage
    -0.60
     understatement
    -0.60
    Fax
    -0.59
     nevertheless
    -0.59
    POSITIVE LOGITS
    romy
    0.92
     mastered
    0.91
    acqu
    0.82
     completes
    0.81
     perfected
    0.79
     finished
    0.78
     completed
    0.78
     sufficiently
    0.78
    sett
    0.78
     stabilized
    0.77
    Activations Density 0.176%

    No Known Activations