© 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. 4-RES-JB
    11. 2794
    Prev
    Next
    INDEX
    Explanations

    It finds words related to the sour taste and sourness

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of sourness

    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
    Constructed
    -0.74
    ITNESS
    -0.73
    atched
    -0.71
    ahon
    -0.70
     Occupations
    -0.68
    Downloadha
    -0.68
    alach
    -0.68
    neapolis
    -0.67
    judicial
    -0.66
     supervised
    -0.66
    POSITIVE LOGITS
     sour
    1.03
     grapes
    0.97
    bean
    0.87
    ashtra
    0.82
    cery
    0.79
    der
    0.79
    cloud
    0.77
    ness
    0.76
    fruit
    0.75
    dy
    0.75
    Activations Density 0.010%

    No Known Activations