© 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 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. 27723
    Prev
    Next
    INDEX
    Explanations

    instances of the word "However" indicating a contrast or exception in the text

    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
    36
    +0.12
    0.7%
    23
    +0.11
    0.7%
    508
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.12
    0.04
    90
    +0.11
    0.04
    508
    +0.10
    0.03
    Negative Logits
    ĨĴ
    -2.16
    Ļª
    -1.61
    ables
    -1.61
    º
    -1.57
    elling
    -1.51
    »¿
    -1.50
    atories
    -1.47
    ļ
    -1.44
     vs
    -1.42
     versus
    -1.41
    POSITIVE LOGITS
    лÑİ
    1.65
    з
    1.65
    tons
    1.61
    Ñīе
    1.51
    омÑĥ
    1.48
    ÑģÑĮ
    1.47
    gni
    1.42
    Ñī
    1.42
    culos
    1.41
    thor
    1.41
    Activations Density 0.052%

    No Known Activations