INDEX

Explanations

occurrences of particular words or phrases likely indicating specific grammatical constructs

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

Followed by Japanese particles の, に, や, か

np_acts-logits-general · gemini-2.0-flash

例のサイトの会員

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

el

-1.30

els

-0.88

大

-0.81

情况下

-0.75

記事

-0.75

大海

-0.73

場合

-0.72

时候

-0.70

特徴

-0.70

the

-0.68

POSITIVE LOGITS



0.69

rungsseite

0.68

IUrlHelper

0.65

 MainAxisSize

0.57

 nakalista

0.55

BeginContext

0.54

 arşivlendi

0.54

 שוליים

0.54

GLI

0.54

 ویکی‌پدیای

0.54

Activations Density 0.033%

occurrences of particular words or phrases likely indicating specific grammatical constructs

Followed by Japanese particles の, に, や, か

例のサイトの会員

No Comments

No Known Activations