© 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. 13768
    Prev
    Next
    INDEX
    Explanations

    terms related to the Core framework, particularly versions and components

    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
    307
    +0.13
    0.7%
    146
    +0.11
    0.6%
    205
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    146
    +0.13
    0.03
    118
    +0.11
    0.02
    379
    +0.11
    0.02
    Negative Logits
     bye
    -1.67
    adays
    -1.60
    ieu
    -1.59
    ancing
    -1.54
    anish
    -1.51
    AY
    -1.51
    ances
    -1.48
    ergic
    -1.47
    agra
    -1.46
     occasion
    -1.46
    POSITIVE LOGITS
    lli
    2.00
    ceptor
    1.83
    quisite
    1.83
    Graphics
    1.77
    ĸ´
    1.77
    ceptors
    1.67
    Foundation
    1.66
    utils
    1.66
    ¿
    1.66
     chemistry
    1.64
    Activations Density 0.170%

    No Known Activations