© 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. 4-RES-SM
    11. 10563
    Prev
    Next
    INDEX
    Explanations

    symbols and punctuation marks

    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
    271
    +0.16
    0.9%
    306
    +0.12
    0.7%
    147
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    129
    +0.16
    0.02
    461
    +0.12
    0.03
    38
    +0.11
    0.03
    Negative Logits
     unnecessary
    -1.45
     based
    -1.36
     tend
    -1.36
     mistake
    -1.35
     proportions
    -1.35
    zona
    -1.34
     illusion
    -1.34
     effect
    -1.33
     least
    -1.32
     proportional
    -1.31
    POSITIVE LOGITS
    ...\...\
    1.97
    ı
    1.76
    xspace
    1.75
    IJ
    1.73
    oddsidemargin
    1.65
    #:
    1.64
    usepackage
    1.57
    dagger
    1.57
    doibase
    1.55
    -\-\
    1.55
    Activations Density 0.062%

    No Known Activations