Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. DeepSeek-R1-Dist-Llama-8B
    3. 15-LLAMASCOPE-SLIMPJ-RES-32K
    4. 1631
    Prev
    Next
    INDEX
    Explanations

    instances of emphasis phrases and qualifiers

    oai_token-act-pair · gpt-4o-mini

    phrases that convey a sense of being limited by only a few factors or conditions.

    oai_token-act-pair · gpt-4oTriggered by @nicholasraphaelgt

    words and phrases indicating quantity or measurement, particularly using numbers like "only", "just", or "as many".

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @emanuelpablo75
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    fnlp/Llama-Scope-R1-Distill/800M-Slimpajama-0-OpenR1-Math-220k/L15R
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Hzfinfdu/SlimPajama-3B
    Features
    32,768
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Hook Layer
    15
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    .way
    -0.17
    ynchronously
    -0.15
    ema
    -0.15
    \<^
    -0.15
    gressor
    -0.15
    ologically
    -0.14
    èµ·æĿ¥
    -0.14
    istrov
    -0.14
    atatype
    -0.14
    é»
    -0.14
    POSITIVE LOGITS
    ableObject
    0.15
     a
    0.15
    deo
    0.14
    ulo
    0.14
    edList
    0.14
     being
    0.14
    ëĿ½
    0.14
     the
    0.14
    Ľå»º
    0.13
    reative
    0.13
    Activations Density 0.257%

    No Known Activations