INDEX

Explanations

instances of deception or disguised intentions

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

making things appear false

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

AndEndTag

-0.54

PMailer

-0.53

RTLU

-0.53

 CreateTagHelper

-0.49

ostomy

-0.48

alist

-0.47

MigrationBuilder

-0.47

 reminder

-0.45

 defaultstate

-0.45

getragen

-0.45

POSITIVE LOGITS

 nonexistent

0.63

 inexist

0.52

 false

0.50

 palsu

0.47

 falso

0.46

 fictitious

0.43

 falsos

0.43

 fake

0.43

 somebody

0.43

 falsa

0.41

Activations Density 0.060%

instances of deception or disguised intentions

making things appear false

No Comments

No Known Activations