© 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. 13214
    Prev
    Next
    INDEX
    Explanations

    important keywords that indicate software configuration and programming concepts

    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
    12
    +0.22
    1.3%
    271
    +0.17
    1.0%
    263
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    345
    +0.22
    0.10
    330
    +0.17
    0.10
    66
    +0.15
    0.13
    Negative Logits
     exception
    -1.63
    )],
    -1.41
     commens
    -1.40
     respectively
    -1.39
     thereof
    -1.36
     exceptions
    -1.36
     average
    -1.35
     seasons
    -1.33
    .+
    -1.33
     derivative
    -1.32
    POSITIVE LOGITS
     Janeiro
    1.97
    ¼
    1.90
    ichte
    1.60
     @
    1.58
    ************************************************************************
    1.56
    endra
    1.49
    openh
    1.49
    amas
    1.49
    ¤
    1.49
     Editor
    1.49
    Activations Density 3.085%

    No Known Activations