© 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. 531
    Prev
    Next
    INDEX
    Explanations

    the word "sure" used as an affirmation or confirmation

    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
    373
    +0.16
    0.7%
    8
    +0.11
    0.5%
    187
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    8
    +0.16
    0.04
    187
    +0.11
    0.03
    581
    +0.10
    0.02
    Negative Logits
    intosh
    -0.72
    çīĪ
    -0.67
    inator
    -0.62
    vernment
    -0.62
    ENC
    -0.62
     contrace
    -0.61
    ENCE
    -0.60
    NING
    -0.60
     resil
    -0.60
    ystem
    -0.59
    POSITIVE LOGITS
    footed
    0.98
    fire
    0.96
    ty
    0.85
     enough
    0.78
     Gamble
    0.74
    ndra
    0.73
    ties
    0.72
    faced
    0.71
    flo
    0.68
    staking
    0.67
    Activations Density 0.067%

    No Known Activations