INDEX

Explanations

phrases indicating the complexity and difficulties of situations

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

already bad, difficult, broken, dead

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

WriteBarrier

-0.32

UpInside

-0.32

 hakim

-0.27

ukan

-0.26

違い

-0.25

jsxFileName

-0.24

 contraire

-0.24

kont

-0.24

ved

-0.24

ler

-0.24

POSITIVE LOGITS

AutoScaleMode

0.71

 الرياضيه

0.70

ViewFeatures

0.64

oredCriteria

0.63

 незавершена

0.63

⟬

0.62

 queſta

0.61

 GetEnumerator

0.60

 poveznice

0.60

$_"

0.59

Activations Density 0.023%

phrases indicating the complexity and difficulties of situations

already bad, difficult, broken, dead

No Comments

No Known Activations