INDEX

Explanations

phrases related to warnings and safety precautions

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

warning and undesirable conditions

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

AndEndTag

-0.40

Matricula

-0.36

Tembelea

-0.34

StateException

-0.34

ArrowToggle

-0.33

JNIEnv

-0.32

шься

-0.32

 lặng

-0.31

 ModelExpression

-0.30

 QUANT

-0.29

POSITIVE LOGITS

EDEFAULT

0.64

)_/¯

0.54

CrossRef

0.50

IVEREF

0.47

trav

0.47

ADELPHIA

0.46

Wade

0.46

 OMITBAD

0.45

ayette

0.45

FormTagHelper

0.45

Activations Density 0.010%

phrases related to warnings and safety precautions

warning and undesirable conditions

No Comments

No Known Activations