© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 7454
    Prev
    Next
    INDEX
    Explanations

    instances of the word "mis" repeated multiple times, likely indicating a focus on detecting words related to mistakes or missteps in the text

    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
    35
    +0.12
    0.5%
    0
    +0.11
    0.5%
    746
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    0
    +0.12
    0.02
    35
    +0.11
    0.02
    251
    +0.10
    0.02
    Negative Logits
    ILA
    -0.67
    INGS
    -0.65
     unto
    -0.64
    eteria
    -0.63
    ¯¯¯¯
    -0.61
     sans
    -0.61
     Destroyer
    -0.60
     Robots
    -0.60
     Ready
    -0.59
    ieri
    -0.59
    POSITIVE LOGITS
    cellaneous
    1.42
    appropri
    1.30
    beh
    1.18
    pelled
    1.10
    behavior
    1.10
    informed
    1.09
    aligned
    1.07
    character
    1.03
    jud
    1.01
    managed
    1.01
    Activations Density 0.014%

    No Known Activations