© 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. 11-RES-JB
    11. 4144
    Prev
    Next
    INDEX
    Explanations

    instances of the word "Ha" occurring with high activation values

    oai_token-act-pair · gpt-3.5-turbo

    mentions of the word "Haaretz."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     multiplier
    -0.64
    pants
    -0.61
    âķIJ
    -0.60
    ments
    -0.60
    states
    -0.59
    mented
    -0.58
     unders
    -0.57
    ij士
    -0.56
    stall
    -0.55
    parts
    -0.55
    POSITIVE LOGITS
    aretz
    1.29
    irst
    1.24
    iku
    1.16
    iley
    1.15
    verty
    1.12
    pless
    1.11
    ifa
    1.10
    user
    1.06
    Ha
    1.04
    ilee
    1.03
    Activations Density 0.014%

    No Known Activations