© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 25689
    Prev
    Next
    INDEX
    Explanations

    specific numeric values or identifiers

    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
    310
    +0.15
    0.8%
    31
    +0.13
    0.7%
    490
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    113
    +0.15
    0.07
    497
    +0.13
    0.01
    444
    +0.13
    0.02
    Negative Logits
     heights
    -1.53
     Dreams
    -1.53
     how
    -1.47
     possible
    -1.39
     imagin
    -1.38
     Heights
    -1.38
    ESULT
    -1.38
     Ways
    -1.36
    born
    -1.32
     conceivable
    -1.29
    POSITIVE LOGITS
    ttes
    1.64
    riber
    1.60
    º
    1.52
    atar
    1.48
    icos
    1.45
    ently
    1.44
    cursors
    1.42
     heads
    1.41
    ľĵ
    1.40
     heck
    1.40
    Activations Density 0.518%

    No Known Activations