INDEX

Explanations

phrases indicating paths or connections between points

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

from ... to

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

 Administrativna

-0.49

expandindo

-0.46

AndEndTag

-0.44

حياته

-0.43

 الرياضيه

-0.43

autorelease

-0.42

engagent

-0.42

geslacht

-0.42

Viited

-0.41

mtx

-0.41

POSITIVE LOGITS

 between

0.68

 pomiędzy

0.66

between

0.63

 mellem

0.62

 między

0.60

 между

0.59

 בין

0.58

Between

0.57

 tussen

0.57

enumii

0.57

Activations Density 0.053%

phrases indicating paths or connections between points

from ... to

No Comments

No Known Activations