© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 27078
    Prev
    Next
    INDEX
    Explanations

    references to matrices and their properties

    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
    156
    +0.21
    1.3%
    376
    +0.12
    0.7%
    196
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    392
    +0.21
    0.02
    272
    +0.12
    0.01
    196
    +0.12
    0.01
    Negative Logits
    woke
    -1.76
     whim
    -1.61
    rav
    -1.60
    rapeut
    -1.55
     particular
    -1.51
     victims
    -1.48
    rapy
    -1.48
     young
    -1.41
     suppose
    -1.40
     deceased
    -1.39
    POSITIVE LOGITS
    ycin
    1.75
    esian
    1.70
    ceral
    1.55
    fonts
    1.50
    illin
    1.50
    itious
    1.49
    istry
    1.48
     improves
    1.45
    asek
    1.45
    rams
    1.44
    Activations Density 0.066%

    No Known Activations