© 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 768
    9. 8-RES_FS768-JB
    11. 722
    Prev
    Next
    INDEX
    Explanations

    legal terms and discussions, specifically related to the death penalty and impeachment

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    159
    +0.18
    0.8%
    393
    +0.14
    0.7%
    116
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    159
    +0.18
    0.20
    95
    +0.14
    0.10
    606
    +0.11
    0.12
    Negative Logits
    aws
    -0.82
    icons
    -0.80
    ARS
    -0.79
    isms
    -0.76
    adle
    -0.74
    ickets
    -0.73
    acers
    -0.72
    agi
    -0.71
    mates
    -0.71
    gans
    -0.70
    POSITIVE LOGITS
     week
    1.13
     particular
    1.08
     trope
    1.04
     article
    1.03
     latest
    1.03
     year
    0.99
     newfound
    0.98
     month
    0.96
     guy
    0.96
     weekend
    0.94
    Activations Density 1.675%

    No Known Activations