Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. MLP Post
    5. 0-MLP-SM
    6. 27266
    Prev
    Next
    INDEX
    Explanations

    instances of the word "survive" and related concepts of endurance and resilience

    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
    376
    +0.21
    1.2%
    148
    +0.13
    0.7%
    413
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    419
    +0.21
    0.01
    254
    +0.13
    0.01
    265
    +0.11
    0.01
    Negative Logits
    unnumbered
    -1.82
     Stay
    -1.54
     Trust
    -1.47
    defer
    -1.43
    hyd
    -1.41
    ubicin
    -1.39
    arn
    -1.39
    unde
    -1.37
    oche
    -1.32
    dere
    -1.31
    POSITIVE LOGITS
    Īĺ
    1.69
    Ļª
    1.67
    ¡
    1.59
    ®
    1.54
    vier
    1.49
    ģ
    1.48
    ictions
    1.47
    guards
    1.47
     beforehand
    1.46
    ĻĤ
    1.45
    Activations Density 0.013%

    No Known Activations