© 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. 4-RES-SM
    11. 14408
    Prev
    Next
    INDEX
    Explanations

    elements related to data structures and attributes in code or markup

    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
    78
    +0.18
    1.0%
    185
    +0.12
    0.7%
    280
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    185
    +0.18
    0.04
    36
    +0.12
    0.06
    255
    +0.12
    0.00
    Negative Logits
    ,’
    -2.73
    '),
    -2.53
    .’
    -2.33
    ”),
    -2.14
    ’,
    -2.10
    ,”
    -2.09
    ,'
    -2.02
     ***
    -1.98
     ',
    -1.96
     [@
    -1.92
    POSITIVE LOGITS
    £
    4.57
    ŀ
    4.51
    Ĵ
    4.47
    ³
    4.42
    ł
    4.38
    ĭ
    4.38
    ª
    4.34
    į
    4.33
    ĩ
    4.33
    Ĺ
    4.32
    Activations Density 0.340%

    No Known Activations