© 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. 14961
    Prev
    Next
    INDEX
    Explanations

    references to media releases or announcements

    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
    21
    +0.18
    1.0%
    9
    +0.14
    0.8%
    135
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    9
    +0.18
    0.03
    386
    +0.14
    0.01
    218
    +0.13
    0.03
    Negative Logits
     necrosis
    -1.54
     helper
    -1.47
    eric
    -1.46
     heads
    -1.38
    hip
    -1.33
    hift
    -1.33
     implement
    -1.31
    ieg
    -1.31
    orta
    -1.28
     blindness
    -1.27
    POSITIVE LOGITS
    Ľ
    3.15
    ¬
    2.98
    Ļ
    2.92
    ased
    2.76
    ½
    2.75
    asing
    2.71
    Ĵ
    2.71
    ł
    2.70
    «
    2.65
    ħ
    2.65
    Activations Density 0.159%

    No Known Activations