INDEX

Explanations

negations

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

negation and contractions

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

Slf

-0.57

 okuyayım

-0.54

 piú

-0.47

 ElementRef

-0.45

 speranza

-0.45

 setId

-0.43

☆☆☆☆☆

-0.43

 Trä

-0.42

 ideen

-0.41

 стрем

-0.41

POSITIVE LOGITS

not

0.75

0.69

 necessarily

0.68

not

0.64

IsNot

0.60

 isnt

0.60

NotBe

0.60

 wont

0.60

 arent

0.58

 aint

0.57

Activations Density 0.020%

negations

negation and contractions

No Comments

No Known Activations