© 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
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. MLP Post
    5. 2-MLP-SM
    6. 25682
    Prev
    Next
    INDEX
    Explanations

    syntax and structure elements in programming languages

    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
    387
    +0.15
    0.8%
    156
    +0.13
    0.7%
    331
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    208
    +0.15
    0.04
    330
    +0.13
    0.04
    331
    +0.12
    0.03
    Negative Logits
    áŁ
    -1.49
    á̝
    -1.48
    pntd
    -1.47
    blogs
    -1.47
    pmed
    -1.44
    leine
    -1.44
    àµį
    -1.43
    medscimonit
    -1.42
    à¯į
    -1.36
    IGHT
    -1.33
    POSITIVE LOGITS
     any
    1.63
    any
    1.53
     Promise
    1.49
     angular
    1.44
     this
    1.42
     asteroid
    1.41
    ality
    1.41
     anything
    1.41
    eness
    1.39
     resemblance
    1.39
    Activations Density 0.191%

    No Known Activations