© 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 49152
    9. 8-RES_FS49152-JB
    11. 41743
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "third" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    195
    +0.12
    0.6%
    339
    +0.12
    0.6%
    216
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    402
    +0.12
    0.03
    339
    +0.12
    0.03
    276
    +0.10
    0.03
    Negative Logits
    staking
    -0.81
    veyard
    -0.66
    immers
    -0.66
    ++++++++++++++++
    -0.66
    UTION
    -0.65
    ITED
    -0.64
     Limits
    -0.64
    values
    -0.63
     Monsters
    -0.62
    Money
    -0.60
    POSITIVE LOGITS
     baseman
    1.15
    rd
    1.02
     inning
    1.01
     installment
    0.95
     consecutive
    0.92
     quarter
    0.83
     iteration
    0.82
     edition
    0.80
    irty
    0.79
    ousand
    0.78
    Activations Density 0.035%

    No Known Activations