© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 4-RES-SM
    6. 11715
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and separators used within the text

    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
    +0.22
    1.2%
    156
    +0.13
    0.7%
    0
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    264
    +0.22
    0.04
    182
    +0.13
    0.04
    353
    +0.12
    0.04
    Negative Logits
    »¿
    -2.08
     respect
    -1.48
     Hole
    -1.38
     regard
    -1.37
    ogi
    -1.34
     rehe
    -1.32
     accompanies
    -1.32
    rophy
    -1.32
    hee
    -1.31
    č↵    
    -1.31
    POSITIVE LOGITS
    blogspot
    2.74
    wordpress
    2.71
    github
    2.57
    org
    2.41
    yahoo
    2.35
    googleapis
    2.30
    apache
    2.25
    mozilla
    2.23
    google
    2.19
    substack
    2.10
    Activations Density 0.097%

    No Known Activations