© 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. 18791
    Prev
    Next
    INDEX
    Explanations

    phrases related to beliefs, thoughts, and hopes

    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
    447
    +0.20
    0.9%
    142
    +0.12
    0.5%
    642
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    142
    +0.20
    0.05
    244
    +0.12
    0.04
    407
    +0.10
    0.04
    Negative Logits
    clock
    -0.79
    rendered
    -0.70
    inary
    -0.67
    aration
    -0.66
    abi
    -0.66
    info
    -0.66
    til
    -0.63
    alia
    -0.63
     Written
    -0.63
    css
    -0.62
    POSITIVE LOGITS
     himself
    0.70
     positives
    0.65
     pacing
    0.61
     herself
    0.61
     passionately
    0.61
     RIS
    0.60
     anecd
    0.59
     sclerosis
    0.59
    paces
    0.59
     optimism
    0.58
    Activations Density 0.229%

    No Known Activations