© 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. MLP Post
    9. 0-MLP-SM
    11. 25943
    Prev
    Next
    INDEX
    Explanations

    occurrences of high activation values indicating emphasis or strong sentiment in the text

    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/0-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.0.hook_mlp_out
    Hook Layer
    0
    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
    223
    +0.13
    0.7%
    408
    +0.13
    0.7%
    310
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    438
    +0.13
    0.40
    53
    +0.13
    0.42
    56
    +0.12
    0.11
    Negative Logits
     irrit
    -1.55
     objection
    -1.46
     iron
    -1.45
    ock
    -1.44
     oxid
    -1.40
     agitated
    -1.39
    IDE
    -1.37
     tick
    -1.34
    ster
    -1.33
     indicate
    -1.33
    POSITIVE LOGITS
    ries
    1.70
    oday
    1.67
    WISE
    1.65
     whatsoever
    1.54
    nai
    1.46
    adata
    1.45
    izations
    1.45
    WHERE
    1.42
    decision
    1.41
    isations
    1.39
    Activations Density 4.688%

    No Known Activations