© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 31809
    Prev
    Next
    INDEX
    Explanations

    instances of phrases related to programming or code structure

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    160
    +0.13
    0.7%
    306
    +0.13
    0.7%
    387
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    75
    +0.13
    0.02
    170
    +0.13
    0.02
    160
    +0.12
    0.03
    Negative Logits
     sense
    -1.77
     judgment
    -1.52
     thereto
    -1.51
    onium
    -1.51
     couple
    -1.48
    theless
    -1.40
    ricanes
    -1.39
     judgments
    -1.39
     induced
    -1.39
     inactivated
    -1.37
    POSITIVE LOGITS
    bows
    1.88
    ¿½
    1.78
    ĻĤ
    1.62
    IJ
    1.60
    eligible
    1.57
    ĵ
    1.54
    ĥ
    1.53
    »¿
    1.52
     Wales
    1.51
    «
    1.47
    Activations Density 0.097%

    No Known Activations