© 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. 23447
    Prev
    Next
    INDEX
    Explanations

    instances where the word "but" is used to introduce contrasting ideas

    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
    38
    +0.15
    0.8%
    448
    +0.13
    0.7%
    405
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    274
    +0.15
    0.01
    247
    +0.13
    0.01
    149
    +0.12
    0.00
    Negative Logits
    resis
    -1.60
    àµį
    -1.52
    hardt
    -1.49
    oqu
    -1.46
    oS
    -1.45
     BGCOLOR
    -1.44
     accordingly
    -1.37
    offs
    -1.34
    ophys
    -1.32
     probabilities
    -1.30
    POSITIVE LOGITS
    inder
    1.50
    watson
    1.44
    aly
    1.40
    pone
    1.32
    ance
    1.31
    orce
    1.27
    gra
    1.27
    cil
    1.27
     permits
    1.27
    ally
    1.26
    Activations Density 0.001%

    No Known Activations