© 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. 4200
    Prev
    Next
    INDEX
    Explanations

    sentences that include the word 'message' and potentially emotional or impactful words

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to messages or themes of communication

    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
     empir
    -0.84
     challeng
    -0.84
     advantageous
    -0.82
     indispensable
    -0.82
     phenomen
    -0.81
     disparate
    -0.79
     inertia
    -0.79
     fundamentally
    -0.78
     explicitly
    -0.78
     synthes
    -0.77
    POSITIVE LOGITS
    ï¸ı
    1.25
    Police
    1.22
    âĢ
    1.18
    Neigh
    1.16
    police
    1.15
    Family
    1.14
    ðŁĺ
    1.12
    Investigators
    1.12
    Authorities
    1.08
    Officers
    1.08
    Activations Density 0.309%

    No Known Activations