© 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. 4-RES-SM
    11. 11907
    Prev
    Next
    INDEX
    Explanations

    HTML tags and attributes related to document 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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    21
    +0.22
    1.3%
    78
    +0.20
    1.2%
    255
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.22
    0.01
    203
    +0.20
    0.04
    78
    +0.13
    0.00
    Negative Logits
    ater
    -1.43
    kins
    -1.40
    rophy
    -1.40
    ator
    -1.39
    estock
    -1.38
    otive
    -1.36
    shaw
    -1.32
    urns
    -1.31
    elson
    -1.26
    umab
    -1.26
    POSITIVE LOGITS
    Ĥ
    3.24
    Ħ
    3.23
    ³
    3.17
    ¸
    3.14
    Ĺ
    3.13
    °
    3.10
    «
    3.08
    ¹
    3.00
    ®
    3.00
    ·
    2.97
    Activations Density 0.181%

    No Known Activations