© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 8-RES-JB
    11. 21061
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical actions involving jumping or plunging downward

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "dive" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     ACTIONS
    -0.68
     Mechdragon
    -0.67
    minus
    -0.67
    iah
    -0.66
    weather
    -0.66
    okers
    -0.63
    abeth
    -0.61
    naires
    -0.61
    addon
    -0.60
    Percent
    -0.60
    POSITIVE LOGITS
     Dive
    1.02
     diving
    1.02
     dive
    1.02
    tails
    0.90
    tail
    0.89
     dives
    0.85
    kick
    0.79
    ulic
    0.79
    ging
    0.77
    river
    0.76
    Activations Density 0.032%

    No Known Activations