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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 25847
    Prev
    Next
    INDEX
    Explanations

    mentions of "people" 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/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
    163
    +0.13
    0.7%
    23
    +0.12
    0.6%
    143
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    38
    +0.13
    0.01
    237
    +0.12
    0.02
    391
    +0.12
    0.03
    Negative Logits
    ¹
    -3.33
    º
    -3.04
    Ľ
    -3.02
    ®
    -3.00
    Īĺ
    -2.95
    Ħ
    -2.79
    ª
    -2.73
    Ŀ
    -2.73
    ĻĤ
    -2.72
    ĭ
    -2.71
    POSITIVE LOGITS
    las
    1.89
    lies
    1.51
    sufficient
    1.50
    fired
    1.48
    orts
    1.47
     lead
    1.46
    ounder
    1.46
    ln
    1.44
    bestos
    1.40
     seat
    1.40
    Activations Density 0.268%

    No Known Activations