© 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. MLP Post
    9. 2-MLP-SM
    11. 22702
    Prev
    Next
    INDEX
    Explanations

    references to data and datasets

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__mlp-sm_processed/2-mlp-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_mlp_out
    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
    419
    +0.13
    0.7%
    455
    +0.12
    0.6%
    121
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    56
    +0.13
    0.03
    120
    +0.12
    0.03
    505
    +0.10
    0.02
    Negative Logits
    ulse
    -1.53
    opause
    -1.48
    ]];
    -1.46
     attention
    -1.41
    cco
    -1.40
     lyrics
    -1.37
     onto
    -1.36
     maybe
    -1.33
    ogon
    -1.32
     status
    -1.32
    POSITIVE LOGITS
    ^--
    1.69
    ^âĪĴ
    1.62
    ayers
    1.54
    isans
    1.51
    igator
    1.51
    feas
    1.44
    idable
    1.38
     Congress
    1.36
     IJ
    1.35
    IAM
    1.33
    Activations Density 0.090%

    No Known Activations