© 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. 2-RES-SM
    11. 32039
    Prev
    Next
    INDEX
    Explanations

    references to programming constructs and 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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    320
    +0.26
    1.5%
    497
    +0.14
    0.8%
    72
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    320
    +0.26
    0.15
    497
    +0.14
    0.05
    187
    +0.13
    0.07
    Negative Logits
    hest
    -1.89
    yi
    -1.69
    idin
    -1.34
    Framework
    -1.31
    enium
    -1.30
    thm
    -1.30
     Framework
    -1.28
    ong
    -1.28
    problem
    -1.27
    ando
    -1.26
    POSITIVE LOGITS
    omed
    1.42
    OLOG
    1.37
     Âĵ
    1.32
     **_
    1.31
    ERIAL
    1.31
    ILE
    1.27
     kne
    1.25
     Morm
    1.23
     osteoporosis
    1.22
     ""
    1.22
    Activations Density 4.096%

    No Known Activations