© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 26413
    Prev
    Next
    INDEX
    Explanations

    expressions of frustration and criticism regarding content

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    258
    +0.24
    1.4%
    23
    +0.20
    1.2%
    56
    +0.19
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    258
    +0.24
    0.21
    23
    +0.20
    0.16
    321
    +0.19
    0.11
    Negative Logits
     himself
    -1.74
     viewpoint
    -1.69
    aine
    -1.69
     quarters
    -1.69
    ème
    -1.62
     aboard
    -1.58
    ural
    -1.58
    tees
    -1.57
    iga
    -1.53
    CoV
    -1.53
    POSITIVE LOGITS
     Hopefully
    1.80
     Maybe
    1.69
     Then
    1.68
    »
    1.67
     Plus
    1.65
     Secondly
    1.65
     Luckily
    1.63
    txt
    1.62
    COM
    1.61
    rap
    1.60
    Activations Density 1.545%

    No Known Activations