INDEX

Explanations

terms related to usefulness or helpfulness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_22/width_16k/average_l0_123

Prompts (Dashboard)

24,576 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

 Houſe

-0.82

 ſtate

-0.78

 Majefty

-0.77

 avoient

-0.74

 purpoſe

-0.72

 houſe

-0.72

 auroit

-0.71

 ſta

-0.69

 feroit

-0.65

 juſ

-0.65

POSITIVE LOGITS

useful

0.54

atious

0.51

ferous

0.50

iferous

0.49

iculous

0.48

ulous

0.47

 Useful

0.47

wonderful

0.47

tuous

0.47

Wonderful

0.47

Activations Density 0.157%

terms related to usefulness or helpfulness

No Comments

No Known Activations