© 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. 10146
    Prev
    Next
    INDEX
    Explanations

    occurrences of words or phrases indicating relationships and interactions

    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
    466
    +0.17
    1.0%
    320
    +0.13
    0.7%
    209
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    466
    +0.17
    0.13
    330
    +0.13
    0.12
    209
    +0.11
    0.10
    Negative Logits
    ĻĤ
    -2.14
    ¾
    -2.06
    »¿
    -1.97
    Ģ
    -1.92
    IJ
    -1.87
    ij
    -1.75
    Ħ
    -1.62
    ĸ
    -1.59
    ·
    -1.58
     meant
    -1.57
    POSITIVE LOGITS
     redistribute
    3.17
     modify
    2.89
     organize
    2.77
     distribute
    2.76
     develop
    2.76
     improve
    2.76
     reduce
    2.70
     cooperate
    2.68
     communicate
    2.66
     eliminate
    2.65
    Activations Density 1.241%

    No Known Activations