© 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 12288
    9. 8-RES_FS12288-JB
    11. 811
    Prev
    Next
    INDEX
    Explanations

    discussion of political and social issues within a specific context

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    442
    +0.11
    0.5%
    430
    +0.10
    0.5%
    757
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    368
    +0.11
    0.03
    726
    +0.10
    0.02
    221
    +0.09
    0.02
    Negative Logits
    bang
    -0.62
     rall
    -0.61
    letes
    -0.58
    reset
    -0.58
     Reincarn
    -0.58
    gged
    -0.57
    itial
    -0.56
     Prin
    -0.56
     Trooper
    -0.56
     Maid
    -0.55
    POSITIVE LOGITS
     relativity
    0.79
    orea
    0.77
    berus
    0.71
    ombat
    0.69
    .--
    0.64
     è£ıè
    0.63
     ILCS
    0.63
    edia
    0.62
    agi
    0.61
    eteria
    0.61
    Activations Density 0.135%

    No Known Activations