INDEX

Explanations

references to race and historical context related to oppression

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

Follows infinitive "to"

np_acts-logits-general · gemini-2.0-flash

to + verb structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

 beschik

-0.78

 écout

-0.72

 feroit

-0.69

 déchir

-0.69

 auroit

-0.68

 ainfi

-0.67

 avoient

-0.67

 démocr

-0.65

 équip

-0.64

 arrêté

-0.64

POSITIVE LOGITS

 romanti

1.01

 legitim

0.90

 glor

0.85

 glorify

0.84

 glam

0.84

 legiti

0.78

glor

0.77

 normalize

0.74

 gloss

0.73

 objec

0.72

Activations Density 0.486%

references to race and historical context related to oppression

Follows infinitive "to"

to + verb structures

No Comments

No Known Activations