© 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. 37216
    Prev
    Next
    INDEX
    Explanations

    phrases that emphasize repetition and redundancy

    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
    660
    +0.12
    0.5%
    526
    +0.12
    0.5%
    98
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    660
    +0.12
    0.01
    98
    +0.12
    0.01
    275
    +0.11
    0.01
    Negative Logits
     Authors
    -0.62
    Ge
    -0.61
    Thirty
    -0.60
    saf
    -0.59
    itans
    -0.57
    vana
    -0.56
    afety
    -0.56
     Syndicate
    -0.56
    VICE
    -0.56
    ogens
    -0.56
    POSITIVE LOGITS
     again
    1.06
    etheless
    0.94
    drive
    0.94
    again
    0.81
    ride
    0.76
    clock
    0.75
     until
    0.73
    repe
    0.73
    stretched
    0.70
    haul
    0.70
    Activations Density 0.009%

    No Known Activations