INDEX

Explanations

programming-related tokens and structures, particularly in a coding or mathematical context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

featureID

-0.74

SourceChecksum

-0.71

 surla

-0.66

httphttps

-0.61

+};

-0.61



-0.59

}]

-0.59

fromnode

-0.57

 otomatig

-0.57

 tartalomajánló

-0.56

POSITIVE LOGITS

 inverted

0.80

 reversed

0.80

 reverse

0.73

 invert

0.68

 inversion

0.66

inverted

0.66

 reversing

0.65

 inverse

0.65

 Reverse

0.65

 inver

0.63

Activations Density 0.896%

programming-related tokens and structures, particularly in a coding or mathematical context

No Comments

No Known Activations