© 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 24576
    9. 8-RES_FS24576-JB
    11. 9364
    Prev
    Next
    INDEX
    Explanations

    news articles related to various incidents and events

    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
    373
    +0.18
    0.8%
    8
    +0.12
    0.6%
    624
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    580
    +0.18
    0.04
    493
    +0.12
    0.03
    8
    +0.11
    0.04
    Negative Logits
    Interstitial
    -0.64
     purity
    -0.61
    Sequ
    -0.59
    reality
    -0.59
    ults
    -0.59
     Subtle
    -0.58
    Viol
    -0.57
    Zone
    -0.56
     occurrence
    -0.55
     Palest
    -0.54
    POSITIVE LOGITS
     tempted
    0.84
    onna
    0.82
     poised
    0.77
     eyeing
    0.76
    iere
    0.74
     headed
    0.70
     aback
    0.69
     guilty
    0.68
    ¶
    0.67
    selling
    0.66
    Activations Density 0.155%

    No Known Activations