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. 27326
    Prev
    Next
    INDEX
    Explanations

    references to the word "Black" in various contexts

    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
    156
    +0.16
    0.9%
    169
    +0.12
    0.7%
    452
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    452
    +0.16
    0.02
    1
    +0.12
    0.02
    506
    +0.12
    0.02
    Negative Logits
    ·¸
    -3.44
    Ĭ
    -3.26
    ĺ
    -3.18
    ´
    -3.17
    ĥ½
    -3.16
    ¬
    -3.12
    Ĺ
    -3.02
    Ń
    -3.01
    ĭ
    -2.99
    Ħ
    -2.90
    POSITIVE LOGITS
    pool
    1.93
    bank
    1.89
    strand
    1.88
    Berry
    1.83
    stown
    1.67
    nature
    1.67
    employment
    1.65
    doll
    1.52
    jack
    1.51
    water
    1.50
    Activations Density 0.040%

    No Known Activations