© 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. 9292
    Prev
    Next
    INDEX
    Explanations

    instances of the word "wrong" and expressions of disagreement or problems

    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
    390
    +0.13
    0.7%
    138
    +0.12
    0.7%
    240
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    403
    +0.13
    0.02
    117
    +0.12
    0.04
    390
    +0.11
    0.04
    Negative Logits
    apine
    -1.65
    chle
    -1.61
    ilities
    -1.57
    oxacin
    -1.54
    FROM
    -1.47
    googleapis
    -1.47
    avir
    -1.42
    ocity
    -1.41
    ]{.
    -1.38
    cyclopedia
    -1.34
    POSITIVE LOGITS
    fully
    2.50
    headed
    2.14
    doing
    1.86
    fulness
    1.80
     Answer
    1.80
     wording
    1.70
    do
    1.65
     answer
    1.65
    ed
    1.63
    acted
    1.60
    Activations Density 0.289%

    No Known Activations