INDEX

Explanations

indented lines or blank spaces, possibly indicating structural elements in a code or document format

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

Followed/preceded by formatting or structural tokens

np_acts-logits-general · gemini-2.0-flash

values and mathematical questions

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

um

-0.78

$',

-0.75

$",

-0.75

ak

-0.73

met

-0.73

len

-0.72

da

-0.71

ir

-0.71

om

-0.71

POSITIVE LOGITS

 itſelf

1.48

 myſelf

1.39

 pleaſure

1.34

 leſs

1.24

 ſche

1.23

 Jefus

1.18

 beſt

1.15

 juſ

1.12

 greateſt

1.12

 purpoſe

1.12

Activations Density 0.285%

indented lines or blank spaces, possibly indicating structural elements in a code or document format

Followed/preceded by formatting or structural tokens

values and mathematical questions

No Comments

No Known Activations