© 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. 11545
    Prev
    Next
    INDEX
    Explanations

    references to geometric or polygonal concepts in programming or mathematical contexts

    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
    497
    +0.13
    0.7%
    50
    +0.13
    0.7%
    214
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    50
    +0.13
    0.05
    497
    +0.13
    0.01
    506
    +0.13
    -0.01
    Negative Logits
    1451450014514500
    -1.67
     Moody
    -1.59
     Crow
    -1.56
     KC
    -1.53
     auditor
    -1.51
     Heaven
    -1.46
    atable
    -1.41
     Hospit
    -1.41
    sid
    -1.40
     Jamie
    -1.39
    POSITIVE LOGITS
     Pol
    3.85
     pol
    3.59
     Poly
    3.45
    Pol
    3.33
    pol
    3.33
    poly
    3.19
    Poly
    3.06
     poly
    3.05
     POL
    2.72
    POL
    2.68
    Activations Density 0.774%

    No Known Activations