© 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 EvalsExportsAPI 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. 12553
    Prev
    Next
    INDEX
    Explanations

    Java class declarations and references to related components within a codebase

    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
    +0.26
    1.5%
    21
    +0.19
    1.1%
    78
    +0.17
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.26
    0.06
    255
    +0.19
    0.00
    504
    +0.17
    0.03
    Negative Logits
     ourselves
    -1.80
    .).
    -1.71
    ).](
    -1.69
    .](
    -1.64
    \'
    -1.62
     \'
    -1.60
    .),
    -1.57
    ¿½
    -1.57
    .);
    -1.56
    \]](
    -1.53
    POSITIVE LOGITS
     {
    2.37
     extends
    2.36
     implements
    2.22
    (){
    2.03
     {};
    2.01
    Ļª
    1.82
    ">
    1.79
     :
    1.78
    TestCase
    1.76
    Impl
    1.74
    Activations Density 0.644%

    No Known Activations