© 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. 4-RES-SM
    11. 14011
    Prev
    Next
    INDEX
    Explanations

    numerical values, particularly those related to legal cases or specific data points

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    156
    +0.39
    2.3%
    255
    +0.12
    0.7%
    477
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    458
    +0.39
    0.03
    477
    +0.12
    0.06
    37
    +0.10
    0.01
    Negative Logits
    itian
    -1.60
     heap
    -1.47
    ible
    -1.44
    gy
    -1.42
     covering
    -1.38
    vas
    -1.37
    matically
    -1.32
    thood
    -1.28
    taire
    -1.28
     machine
    -1.28
    POSITIVE LOGITS
    ²
    3.17
    Ļ
    3.02
    ·¸
    2.96
    ³
    2.90
    ł
    2.85
    ĻĤ
    2.77
    ¨
    2.67
    »
    2.62
    Ĥ¬
    2.61
    Ļª
    2.56
    Activations Density 0.475%

    No Known Activations