© 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. 1-RES-SM
    11. 26224
    Prev
    Next
    INDEX
    Explanations

    patterns or structures related to programming constructs or code syntax

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    165
    +0.13
    0.7%
    491
    +0.12
    0.6%
    410
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    165
    +0.13
    0.08
    374
    +0.12
    0.11
    312
    +0.11
    0.08
    Negative Logits
    Ļª
    -1.84
    ĩ
    -1.51
     Injury
    -1.45
    rive
    -1.44
     Register
    -1.41
    »¿
    -1.41
     Mut
    -1.34
    ¹
    -1.33
    ulators
    -1.32
    ·¸
    -1.32
    POSITIVE LOGITS
     subjects
    1.67
     astr
    1.42
     disciples
    1.38
     committed
    1.36
    %.
    1.32
     spokes
    1.29
     figures
    1.28
    -----
    1.26
    ingly
    1.25
     matter
    1.25
    Activations Density 0.920%

    No Known Activations