© 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. 2-MLP-SM
    11. 23928
    Prev
    Next
    INDEX
    Explanations

    phrases indicating hierarchical or authority structures

    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
    112
    +0.14
    0.7%
    303
    +0.12
    0.7%
    459
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    112
    +0.14
    0.05
    219
    +0.12
    0.04
    244
    +0.10
    0.03
    Negative Logits
    erl
    -1.39
    ģ
    -1.30
     onto
    -1.29
     inherit
    -1.27
    elen
    -1.27
     onward
    -1.26
     Copyright
    -1.25
     Welfare
    -1.25
    \]](
    -1.21
     belie
    -1.21
    POSITIVE LOGITS
    ”:
    1.67
    blockList
    1.67
    ”
    1.65
    ”.
    1.64
    ”,
    1.61
    ”—
    1.57
    genic
    1.54
    ”?
    1.49
     fois
    1.49
    veolar
    1.47
    Activations Density 0.204%

    No Known Activations