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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 3-RES-JB
    6. 18979
    Prev
    Next
    INDEX
    Explanations

    words related to violence and injuries, especially those related to blood

    oai_token-act-pair · gpt-3.5-turbo

    mentions of blood and related imagery

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Amend
    -0.77
    awaru
    -0.77
     VIDEOS
    -0.77
    ECH
    -0.77
    OPLE
    -0.76
    IX
    -0.76
    AMA
    -0.76
    ALS
    -0.74
    Published
    -0.73
    Subscribe
    -0.72
    POSITIVE LOGITS
    thirst
    1.20
    bath
    1.13
    stained
    1.09
    hound
    1.04
     vessels
    0.98
    wine
    0.97
    lust
    0.96
     thirsty
    0.93
    shed
    0.92
     clot
    0.87
    Activations Density 0.013%

    No Known Activations