© 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. 1452
    Prev
    Next
    INDEX
    Explanations

    advice-related phrases

    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
    679
    +0.17
    0.8%
    101
    +0.11
    0.5%
    149
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    679
    +0.17
    0.04
    247
    +0.11
    0.02
    217
    +0.09
    0.02
    Negative Logits
     Alive
    -0.68
    BP
    -0.67
    plex
    -0.66
    bang
    -0.64
     Canad
    -0.61
     hangar
    -0.61
     Mississ
    -0.61
     Craw
    -0.61
     spect
    -0.61
    itol
    -0.61
    POSITIVE LOGITS
     Advice
    1.17
     advice
    1.16
     advising
    1.09
     advise
    1.01
     advised
    0.93
     counselor
    0.86
     advises
    0.85
     recommending
    0.83
     guidance
    0.82
    arial
    0.81
    Activations Density 0.118%

    No Known Activations