INDEX

Explanations

instances of structured data or programming constructs

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

ſelf

-1.02

transQ

-0.88

 faſt

-0.85

 queſta

-0.85

NameInMap

-0.78



-0.78

 nahilalakip

-0.78

ſelves

-0.78

 ujednoznacz

-0.77

 endforeach

-0.76

POSITIVE LOGITS

the

0.59

The

0.56

if

0.45

The

0.43

the

0.41

 chegada

0.38

After

0.36

If

0.36

Either

0.35

Activations Density 0.141%

instances of structured data or programming constructs

No Comments

No Known Activations