© 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 3072
    9. 8-RES_FS3072-JB
    11. 2967
    Prev
    Next
    INDEX
    Explanations

    phrases related to proving or demonstrating something

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    30
    +0.11
    0.5%
    724
    +0.10
    0.4%
    377
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    724
    +0.11
    0.04
    30
    +0.10
    0.03
    707
    +0.09
    0.04
    Negative Logits
     newsletters
    -0.80
    letal
    -0.80
    umbn
    -0.77
    arta
    -0.72
    adish
    -0.71
    lished
    -0.67
    ades
    -0.67
    ataka
    -0.66
    yip
    -0.66
    rompt
    -0.64
    POSITIVE LOGITS
    ance
    0.76
    reader
    0.75
     incapable
    0.73
     untrue
    0.69
     decisive
    0.69
    worthiness
    0.69
    ||||
    0.68
    manship
    0.68
     resilient
    0.68
     ineffective
    0.67
    Activations Density 0.409%

    No Known Activations