INDEX

Explanations

lines or sections that serve as dividers or separators in the text

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

Paris

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

aarrggbb

-0.99

IntoConstraints

-0.85

 AssemblyCompany

-0.75

<unused52>

-0.73

<unused41>

-0.73

<unused28>

-0.73

[@BOS@]

-0.73

RectangleBorder

-0.72

<unused17>

-0.72

<unused14>

-0.72

POSITIVE LOGITS

 nakalista

0.31

 París

0.28

Extras

0.26

 ничего

0.25

pè

0.25

egz

0.24

 Vielen

0.23

 Flüs

0.23

XCT

0.23

WindowConstants

0.22

Activations Density 0.001%

lines or sections that serve as dividers or separators in the text

Paris

No Comments

No Known Activations