INDEX

Explanations

instances of the word "are" along with variations of "re" indicating second-person references

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

are you

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

 probably

-0.44

probably

-0.41

 quite

-0.39

 Probably

-0.36

 Quite

-0.36

又

-0.35

quite

-0.35

UnsafeEnabled

-0.34

♩

-0.33

Quite

-0.33

POSITIVE LOGITS

 anything

0.84

 jemals

0.81

anything

0.78

 lucky

0.74

Anything

0.71

 yoksa

0.69

 ANYTHING

0.69

 Anything

0.66

 apapun

0.66

 quelconque

0.64

Activations Density 0.020%

instances of the word "are" along with variations of "re" indicating second-person references

are you

No Comments

No Known Activations