INDEX

Explanations

phrases related to problem solving and hints

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

figure

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

ChildScrollView

-0.56

chappement

-0.40

้งาน

-0.35

 manifestó

-0.32

sra

-0.31

 aprobó

-0.30

 onData

-0.29

scheid

-0.29

 temáticas

-0.29

ódź

-0.29

POSITIVE LOGITS

 betweenstory

0.64

Personendaten

0.60

 nakalista

0.59

RTGC

0.56

帖最后由

0.54

 للمعارف

0.52

pecabe

0.52

fromnode

0.51

 guessed

0.49

 guessing

0.47

Activations Density 0.024%

phrases related to problem solving and hints

figure

No Comments

No Known Activations