INDEX

Explanations

phrases indicating a comparative degree or relativity

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

relatively followed by descriptors

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

 your

-0.33

expiration

-0.32

In

-0.31

 ouro

-0.31

Vors

-0.31

onSearch

-0.31

Font

-0.31

pr

-0.30

ൻ

-0.30

action

-0.30

POSITIVE LOGITS

 relatively

1.07

 Relatively

1.01

relatively

0.95

 fairly

0.94

 Fairly

0.88

 comparatively

0.81

 relativ

0.78

 relativement

0.76

OGND

0.76

 Quasi

0.76

Activations Density 0.013%

phrases indicating a comparative degree or relativity

relatively followed by descriptors

No Comments

No Known Activations