INDEX

Explanations

phrases expressing tendencies or inclinations

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

tend to be or do

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

nię

-0.45

 miljö

-0.42

//

-0.41

Comple

-0.41

 escar

-0.40

erapa

-0.40

 deko

-0.40

ailles

-0.40

Substitution

-0.40

 Diſ

-0.40

POSITIVE LOGITS

 tend

1.19

 tends

1.16

tend

0.97

 Tend

0.94

Tend

0.94

 tended

0.93

 tendency

0.91

 tienden

0.87

 tending

0.87

 TEND

0.87

Activations Density 0.011%

phrases expressing tendencies or inclinations

tend to be or do

No Comments

No Known Activations