INDEX

Explanations

requests for advice or suggestions

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

preceding possible solutions or suggestions

np_acts-logits-general · gemini-2.0-flash

asking for how and what

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

 resourceCulture

-0.65

rawDesc

-0.61

BeginContext

-0.57

posables

-0.55

Browsable

-0.53

LookAnd

-0.52

WebElementEntity

-0.52

 AppCompatTheme

-0.51

setopt

-0.51

AddTagHelper

-0.51

POSITIVE LOGITS

how

1.55

how

1.26

 what

1.19

what

1.08

 cómo

1.01

 hvordan

1.00

How

1.00

 bagaimana

0.99

HOW

0.99

why

0.99

Activations Density 0.454%

requests for advice or suggestions

preceding possible solutions or suggestions

asking for how and what

No Comments

No Known Activations