© 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. 0-RES-JB
    11. 19975
    Prev
    Next
    INDEX
    Explanations

    the word "Bergdahl"

    oai_token-act-pair · gpt-3.5-turbo

    repeated mentions of a specific individual's name

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     tyres
    -0.74
     organised
    -0.74
     tier
    -0.71
     PS
    -0.70
     tyre
    -0.69
     upstream
    -0.64
     Premier
    -0.63
     mile
    -0.63
     vector
    -0.62
     plate
    -0.62
    POSITIVE LOGITS
    ahl
    4.47
     Dahl
    1.34
    ahn
    1.21
    orf
    1.16
    igm
    1.11
    ohl
    1.08
    alk
    1.02
    omez
    1.00
    alsh
    1.00
    hower
    0.97
    Activations Density 0.020%

    No Known Activations