© 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. 10-RES-JB
    11. 12205
    Prev
    Next
    INDEX
    Explanations

    phrases or questions related to a specific entity or person indicated by the word "Who"

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "who."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     beginning
    -0.70
     drying
    -0.68
     Grail
    -0.67
     Ground
    -0.63
     yielding
    -0.62
     fixation
    -0.61
     reinforcement
    -0.61
     relaxation
    -0.61
     510
    -0.61
     Supplemental
    -0.60
    POSITIVE LOGITS
    oping
    1.38
    ever
    1.27
    oped
    1.24
    soever
    1.24
    ppers
    1.12
    pper
    1.06
    abouts
    1.01
    ops
    1.00
    osh
    1.00
    zbollah
    1.00
    Activations Density 0.052%

    No Known Activations