INDEX

Explanations

phrases that indicate different perspectives or viewpoints

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

different perspectives or lenses

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

PMailer

-0.53

 kasarigan

-0.46

 للاسماء

-0.45

 demikian

-0.43

SBATCH

-0.43

 myſelf

-0.42

 coû

-0.42

Източници

-0.41

GEBRA

-0.40

 autorytatywna

-0.40

POSITIVE LOGITS

 perspective

0.97

 viewpoint

0.96

 view

0.91

 perspectives

0.89

 viewpoints

0.87

 views

0.85

 PERSPECT

0.77

perspective

0.75

 VIEW

0.75

 View

0.73

Activations Density 0.044%

phrases that indicate different perspectives or viewpoints

different perspectives or lenses

No Comments

No Known Activations