INDEX

Explanations

verbs and actions indicating processes or actions being executed

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

action tokens followed by context

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

enumi

-0.56

 تضيفلها

-0.54

principalTable

-0.51

GHIJKLM

-0.50

AndEndTag

-0.50

MigrationBuilder

-0.50

virons

-0.49

 onCreateView

-0.48

ẵng

-0.47

enumii

-0.47

POSITIVE LOGITS

 Meksiku

0.49

ctez

0.40

تقاوى

0.39

 мәкал

0.37

 autorytatywna

0.36

PET

0.35

sü

0.35

enerbahçe

0.35

nEnter

0.35

 Mulan

0.34

Activations Density 0.171%

verbs and actions indicating processes or actions being executed

action tokens followed by context

No Comments

No Known Activations