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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 4-RES-JB
    6. 18313
    Prev
    Next
    INDEX
    Explanations

    words related to misconduct, corruption, transgressions, and deception

    oai_token-act-pair · gpt-3.5-turbo

    terms related to negative implications or consequences of actions

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

    No Comments

    Negative Logits
    eport
    -0.77
    paren
    -0.77
    sav
    -0.74
    cffff
    -0.70
    soDeliveryDate
    -0.70
    nis
    -0.70
    archives
    -0.68
    raltar
    -0.65
    rians
    -0.65
    ENN
    -0.65
    POSITIVE LOGITS
    ously
    0.87
     havoc
    0.81
    ous
    0.79
    uous
    0.77
     inflicted
    0.74
     perpetrated
    0.73
     Bastard
    0.72
     plag
    0.66
    oxide
    0.65
    Ö¼
    0.63
    Activations Density 0.096%

    No Known Activations