© 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. 13999
    Prev
    Next
    INDEX
    Explanations

    technical terms and syntax related to programming and data structures

    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
    +0.23
    1.4%
    78
    +0.17
    1.0%
    255
    +0.16
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.23
    0.02
    78
    +0.17
    0.03
    185
    +0.16
    0.08
    Negative Logits
    ĥ½
    -2.42
    £
    -2.27
    ↵↵
    -2.19
    ↵
    -2.19
    -2.19
     ↵
    -2.19
    ↵  Âł
    -2.19
     ↵
    -2.19
    <|outofrange|>
    -2.19
    -2.19
    POSITIVE LOGITS
    +"
    2.23
    ULL
    2.00
     +
    1.71
    *(
    1.68
    _);
    1.65
    ++;
    1.63
    *((
    1.59
    +(
    1.59
    );
    1.58
    *(-
    1.57
    Activations Density 1.278%

    No Known Activations