© 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. 2-RES-SM
    11. 10025
    Prev
    Next
    INDEX
    Explanations

    references to specific locations or geographic indicators

    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
    422
    +0.21
    1.3%
    56
    +0.15
    0.9%
    294
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    422
    +0.21
    0.08
    56
    +0.15
    -0.04
    145
    +0.13
    0.13
    Negative Logits
    etheless
    -1.53
    nis
    -1.48
    iamo
    -1.40
     hep
    -1.34
     lightly
    -1.30
     sides
    -1.28
     like
    -1.27
    astically
    -1.26
    handed
    -1.26
    aste
    -1.25
    POSITIVE LOGITS
    ]):
    1.40
    DLL
    1.37
     ()
    1.32
    ooting
    1.31
    FPar
    1.30
    ledge
    1.25
     habit
    1.24
    "}](#
    1.22
    rup
    1.22
    uitar
    1.21
    Activations Density 3.297%

    No Known Activations