INDEX

Explanations

metadata tags in a structured document or code snippet

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

Hexadecimal representations or symbols

np_acts-logits-general · gemini-2.0-flash

otherwise

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-0.65

-0.56

-0.53

↵↵

-0.52

<eos>

-0.52

 userDao

-0.50

and

-0.48

’

-0.46

 highly

-0.43

POSITIVE LOGITS

AnchorStyles

1.22

AddTagHelper

1.20

ImageContext

1.08

mybatisplus

1.06

Personendaten

1.05

UrlResolution

1.05

 '\\;'

1.04

 AssemblyCulture

1.03

InSection

1.02

featureID

1.01

Activations Density 0.384%

metadata tags in a structured document or code snippet

Hexadecimal representations or symbols

otherwise

No Comments

No Known Activations