INDEX

Explanations

repeated use of the word "can" across various contexts

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

"can" followed by a verb

np_acts-logits-general · gemini-2.0-flash

can followed by a verb

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

vician

-0.42

́

-0.41

windowFixed

-0.40

 قاد

-0.40

offer

-0.39

WarningLevel

-0.38

فات

-0.38

œ

-0.38

allow

-0.38

Provide

-0.37

POSITIVE LOGITS

 chi̍t

0.97

 متعلقه

0.97

 expect

0.96

 BoxFit

0.92

 myſelf

0.91

 uſe

0.91

verwijspagina

0.90

 GenerationType

0.88

 ſind

0.86

 Réponses

0.86

Activations Density 0.186%

repeated use of the word "can" across various contexts

"can" followed by a verb

can followed by a verb

No Comments

No Known Activations