© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 7141
    Prev
    Next
    INDEX
    Explanations

    technical terms related to web development and technology

    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
    367
    +0.10
    0.5%
    337
    +0.10
    0.4%
    182
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    76
    +0.10
    0.03
    288
    +0.10
    0.04
    527
    +0.09
    0.02
    Negative Logits
    etheless
    -0.84
     âĵĺ
    -0.68
    rone
    -0.63
    ntil
    -0.62
    intent
    -0.61
    âķIJ
    -0.60
     besides
    -0.58
    cknow
    -0.57
     decency
    -0.56
     whatsoever
    -0.56
    POSITIVE LOGITS
     progresses
    1.29
     unfolds
    1.11
     grows
    1.08
     prepares
    1.06
     evolves
    1.05
     progressed
    0.99
     unfolded
    0.94
     continues
    0.94
     expands
    0.90
     matured
    0.88
    Activations Density 0.256%

    No Known Activations