INDEX

Explanations

terms indicating rank or superiority, particularly the word "top" and its variations

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

top followed by rank or descriptor

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

 זאת

-0.38

 છે

-0.38

-0.37

icleta

-0.36

 Balkon

-0.35

-0.34

existence

-0.34

дли

-0.34

 sarili

-0.34

POSITIVE LOGITS

top

0.91

Top

0.82

TOP

0.79

getTop

0.78

Top

0.73

 Houſe

0.70

 houſe

0.69

 leaſt

0.68

 टॉप

0.66

 beſt

0.65

Activations Density 0.010%

terms indicating rank or superiority, particularly the word "top" and its variations

top followed by rank or descriptor

No Comments

No Known Activations