INDEX

Explanations

instances of the word "same" or variations thereof

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

same followed by qualifier

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

popo

-0.47

PDC

-0.47

Uro

-0.46

 hackers

-0.45

 palanca

-0.44

 povol

-0.44

PPC

-0.44

κος

-0.43

BPI

-0.43

fptr

-0.43

POSITIVE LOGITS

 same

1.33

Same

1.32

 Same

1.23

same

1.20

 SAME

1.08

SAME

1.07

zelfde

0.92

isSame

0.88

 mesmas

0.85

 samma

0.81

Activations Density 0.040%

instances of the word "same" or variations thereof

same followed by qualifier

No Comments

No Known Activations