INDEX

Explanations

terms related to seeking help or guidance

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

Comes before a conditional "if"

np_acts-logits-general · gemini-2.0-flash

helpful suggestions or questions

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

when

-0.79

Unless

-0.72

WHEN

-0.68

should

-0.68

unless

-0.67

whenever

-0.65

 Unless

-0.64

When

-0.61

Should

-0.61

 caso

-0.60

POSITIVE LOGITS

 للمعارف

1.19

 iſt

1.05

ViewFeatures

1.03

 esternos

1.00

 وتسجيلات

0.99

 <<<<<<<<<<<<<<

0.99

 estekak

0.98

 itſelf

0.97

PerformLayout

0.95

 تانيه

0.91

Activations Density 0.227%

terms related to seeking help or guidance

Comes before a conditional "if"

helpful suggestions or questions

No Comments

No Known Activations