INDEX

Explanations

technical terms and key relational phrases in formal contexts

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

numbers related to measurement or counting

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

stdc

-0.36

vœ

-0.32

 ihan

-0.31

***

-0.31

uteen

-0.30

stånd

-0.28

ao

-0.28

so

-0.28

bad

-0.28

owi

-0.28

POSITIVE LOGITS

ſſung

0.68

⟬

0.68

PerformLayout

0.66

<unused8>

0.65

hyrchwyd

0.65

<unused52>

0.64

<unused79>

0.64

<unused41>

0.64

<pad>

0.64

<unused23>

0.64

Activations Density 0.029%

technical terms and key relational phrases in formal contexts

numbers related to measurement or counting

No Comments

No Known Activations