INDEX

Explanations

the word "never" in various forms and contexts

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

never followed by verb/adjective

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

AOC

-0.55

AOC

-0.51

 isSelected

-0.50

 roleId

-0.49

 متحده

-0.49

bufio

-0.48

IMC

-0.48

DCM

-0.48

AOS

-0.48

horizontalLayout

-0.47

POSITIVE LOGITS

 never

1.09

 NEVER

1.03

 Never

1.01

never

0.97

Never

0.97

NEVER

0.95

nunca

0.81

 ever

0.78

 always

0.77

 Nunca

0.77

Activations Density 0.014%

the word "never" in various forms and contexts

never followed by verb/adjective

No Comments

No Known Activations