© 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. 1776
    Prev
    Next
    INDEX
    Explanations

    statements related to feeling comfortable with a certain decision or situation

    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
    288
    +0.13
    0.6%
    710
    +0.10
    0.5%
    578
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    737
    +0.13
    0.01
    710
    +0.10
    0.01
    201
    +0.10
    0.01
    Negative Logits
    bard
    -0.71
    ework
    -0.70
    vernment
    -0.67
    ilee
    -0.65
    grave
    -0.65
    fleet
    -0.65
    hn
    -0.64
    ongyang
    -0.64
    vous
    -0.63
    alez
    -0.63
    POSITIVE LOGITS
     enough
    0.93
    ably
    0.91
     comfortable
    0.86
    fortable
    0.80
    Medium
    0.79
     knowing
    0.73
     indoors
    0.71
     riding
    0.70
     snug
    0.70
     wearing
    0.70
    Activations Density 0.025%

    No Known Activations