INDEX

Explanations

phrases indicating negation or opposition

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ditangkap

-0.59

 Visited

-0.58

 licked

-0.58

converted

-0.57

 sitter

-0.56

branded

-0.56

Pautan

-0.56

ALLOWED

-0.56

StructEnd

-0.56

POSITIVE LOGITS

 EconPapers

0.69

 being

0.65

;"></

0.61

 مرئيه

0.56

izing

0.56

kmäler

0.55

 getting

0.55

 making

0.55

})();

0.54

encodeWith

0.53

Activations Density 0.416%

phrases indicating negation or opposition

No Comments

No Known Activations