INDEX

Explanations

numbers or numerical data in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 assisting

-0.52

 assist

-0.51

 Latimer

-0.45

BibitemOpen

-0.44

 Swain

-0.43

 ajedrez

-0.43

 blij

-0.43

 assistance

-0.42

assist

-0.41

 Larsen

-0.41

POSITIVE LOGITS

1.09

+#+#

0.68

PointerException

0.66

 '\\;'

0.60

########.

0.60

omitempty

0.60

Hundred

0.59

０

0.59

۰

0.59

 Hundred

0.58

Activations Density 0.374%

numbers or numerical data in various contexts

No Comments

No Known Activations