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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 12-RES-JB
    11. 3933
    Prev
    Next
    INDEX
    Explanations

    Explanation of neuron 4 behavior: the main thing this neuron does is find demeaning or insulting words or phrases

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    instances of derogatory or demeaning language

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 12-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     srfAttach
    -0.71
    WP
    -0.69
    GD
    -0.65
    QL
    -0.64
    HT
    -0.64
    rays
    -0.63
    ivation
    -0.62
    rology
    -0.62
    ridor
    -0.60
    REF
    -0.60
    POSITIVE LOGITS
    zz
    0.71
    ber
    0.68
    cl
    0.65
    aning
    0.64
    dropping
    0.63
    aco
    0.61
    farious
    0.60
    bered
    0.58
    back
    0.58
    liner
    0.58
    Activations Density 0.191%

    No Known Activations