© 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. 1-RES-SM
    11. 25645
    Prev
    Next
    INDEX
    Explanations

    specific patterns in code, particularly involving file inclusion and define statements

    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
    478
    +0.12
    0.7%
    101
    +0.12
    0.7%
    149
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.12
    0.06
    355
    +0.12
    0.05
    17
    +0.12
    0.02
    Negative Logits
    ...](
    -1.87
    hower
    -1.79
    ?’
    -1.76
    ?”
    -1.69
    cks
    -1.66
    NEY
    -1.65
    âĢ²
    -1.60
    ERY
    -1.60
    ahi
    -1.59
    ”?
    -1.58
    POSITIVE LOGITS
    µ
    3.79
    ĥ½
    3.79
    ¸
    3.65
    ©
    3.55
    ³
    3.54
    ·
    3.53
    ĨĴ
    3.51
    ·¸
    3.50
    ½
    3.47
    3.45
    Activations Density 0.348%

    No Known Activations