INDEX

Explanations

instances of the word "could" in a variety of contexts

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

"could" followed by verbs

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

<bos>

-0.41

the

-0.32

isFetching

-0.29

 advances

-0.28

<strong>

-0.28

in

-0.28

 Maß

-0.28

必要がある

-0.28

POSITIVE LOGITS

 canst

1.18

could

1.16

 could

1.16

Could

1.14

 COULD

1.13

 Could

1.09

<unused41>

1.02

<unused14>

1.01

<unused8>

1.01

[@BOS@]

1.01

Activations Density 0.039%

instances of the word "could" in a variety of contexts

"could" followed by verbs

No Comments

No Known Activations