INDEX

Explanations

symbolic references or elements denoted by asterisks

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

academic publications

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

Diweddarwch

-0.69

principalColumn

-0.68

 MainAxisSize

-0.61

+#+#

-0.60

bitField

-0.60

RectangleBorder

-0.59

ransition

-0.54

TagMode

-0.54

IntoConstraints

-0.54

InputTagHelper

-0.53

POSITIVE LOGITS

 magazine

0.57

magazine

0.48

 newspaper

0.47

 editors

0.47

 magazines

0.47

 tabloid

0.45

magaz

0.44

 Magazine

0.43

 newspapers

0.42

 reporter

0.42

Activations Density 0.006%

symbolic references or elements denoted by asterisks

academic publications

No Comments

No Known Activations