INDEX

Explanations

the concept of negation or avoidance expressed through the word "never."

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

never followed by verb

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

rinfo

-0.57

DCM

-0.56

bufio

-0.55

royd

-0.54

Pyx

-0.54

horizontalLayout

-0.53

AOC

-0.51

DCM

-0.51

ganu

-0.50

CIE

-0.49

POSITIVE LOGITS

 never

1.70

never

1.62

Never

1.55

 Never

1.55

 NEVER

1.51

NEVER

1.47

nunca

1.24

 jamás

1.22

 Nunca

1.21

 nunca

1.20

Activations Density 0.015%

the concept of negation or avoidance expressed through the word "never."

never followed by verb

No Comments

No Known Activations