© 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. 25825
    Prev
    Next
    INDEX
    Explanations

    terms related to normalization and structuring of data

    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
    156
    +0.27
    1.5%
    162
    +0.17
    1.0%
    148
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    162
    +0.27
    0.07
    148
    +0.17
    0.05
    22
    +0.15
    0.04
    Negative Logits
    lessly
    -2.04
    isans
    -1.85
    fully
    -1.72
    stood
    -1.57
     varied
    -1.50
    rend
    -1.49
    llo
    -1.47
     progressed
    -1.44
    gone
    -1.42
    als
    -1.42
    POSITIVE LOGITS
    river
    1.65
    thon
    1.65
    irectory
    1.60
    noreply
    1.49
    GE
    1.42
    iction
    1.41
     (\>
    1.41
     by
    1.40
    elta
    1.39
     version
    1.38
    Activations Density 0.366%

    No Known Activations