INDEX

Explanations

mentions of numerical data or statistics related to performance indicators

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

bullet points

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

Tikang

-0.67

ILayout

-0.65

arket

-0.62

ArgumentParser

-0.62

ARANCE

-0.61

nastics

-0.60

ooplankton

-0.60

orianCalendar

-0.60

 onix

-0.60

 ecosport

-0.59

POSITIVE LOGITS

•

1.10

••

0.96

•••

0.90

•

0.87

.•

0.83

••••

0.82

)•

0.75

~•

0.74

 ••

0.69

°•

0.66

Activations Density 0.009%

mentions of numerical data or statistics related to performance indicators

bullet points

No Comments

No Known Activations