INDEX

Explanations

phrases that express repetition or continuous action

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

X and X

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

Autoritní

-0.77

 Paglinawan

-0.76

WriteTagHelper

-0.74

rungsseite

-0.68

 Italijanski

-0.67

Билгалдахарш

-0.67

ScopeManager

-0.67

 noDo

-0.66

 propOrder

-0.64

Personendaten

-0.64

POSITIVE LOGITS

 continuamente

0.43

 endlessly

0.41

 banget

0.41

0.40

…

0.38

一个个

0.37

 endless

0.36

0.35

 continuously

0.35

Activations Density 0.007%

phrases that express repetition or continuous action

X and X

No Comments

No Known Activations