© 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. 10506
    Prev
    Next
    INDEX
    Explanations

    terms related to legal claims and court rulings

    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
    168
    +0.20
    1.1%
    138
    +0.17
    1.0%
    251
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    168
    +0.20
    0.25
    138
    +0.17
    0.20
    251
    +0.12
    0.17
    Negative Logits
    ugs
    -1.61
     Architecture
    -1.56
    érie
    -1.56
    rolet
    -1.54
    idian
    -1.48
     favourite
    -1.46
     Growth
    -1.46
     flask
    -1.45
     tuned
    -1.45
     Optim
    -1.45
    POSITIVE LOGITS
     remand
    3.23
     appellate
    3.16
     pleadings
    3.11
     courts
    2.97
     statutory
    2.93
     court
    2.93
     evidentiary
    2.91
    ĸ
    2.88
    FFIRMED
    2.88
     pretrial
    2.88
    Activations Density 4.435%

    No Known Activations