© 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. 9800
    Prev
    Next
    INDEX
    Explanations

    phrases related to the interpretation and analysis of various topics, such as texts, law, and situations

    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
    178
    +0.16
    0.7%
    7
    +0.15
    0.7%
    124
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    178
    +0.16
    0.03
    7
    +0.15
    0.02
    540
    +0.11
    0.02
    Negative Logits
    ccoli
    -0.87
     Indust
    -0.75
    Interstitial
    -0.72
     Empires
    -0.70
    trap
    -0.69
    etsk
    -0.67
    culosis
    -0.67
     Rats
    -0.67
    starter
    -0.66
    nih
    -0.65
    POSITIVE LOGITS
     Interpret
    1.09
     interpret
    1.09
     interpretation
    1.08
     interpretations
    1.06
     meanings
    1.02
    notations
    0.94
    interpret
    0.93
     interpreting
    0.93
     interpre
    0.91
     ambig
    0.91
    Activations Density 0.098%

    No Known Activations