© 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. 2-RES-SM
    11. 10901
    Prev
    Next
    INDEX
    Explanations

    document titles and headings in HTML structure

    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
    55
    +0.12
    0.6%
    38
    +0.11
    0.6%
    394
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    379
    +0.12
    0.02
    38
    +0.11
    0.01
    55
    +0.10
    0.02
    Negative Logits
    CHANTABILITY
    -1.55
     whole
    -1.54
    ?”
    -1.49
    áĢº
    -1.45
    ?).
    -1.44
    oxin
    -1.41
     electricity
    -1.41
     heaven
    -1.40
    fected
    -1.38
     deaths
    -1.38
    POSITIVE LOGITS
    marks
    2.16
    ings
    1.83
    ivities
    1.73
    ivos
    1.69
    ongs
    1.66
    scape
    1.65
     =>
    1.60
    docs
    1.60
    izers
    1.57
    igraphy
    1.55
    Activations Density 0.051%

    No Known Activations