INDEX

Explanations

words related to winning and success

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PDR

-0.52

byshire

-0.51

geot

-0.50

 Giles

-0.50

abase

-0.48

StyleSheet

-0.48

 Magh

-0.48

SDF

-0.48

SH

-0.47

CRL

-0.47

POSITIVE LOGITS

won

0.97

 winner

0.90

 winning

0.89

win

0.88

Won

0.86

winning

0.86

Winning

0.85

won

0.85

 WINNER

0.83

 Winning

0.82

Activations Density 0.062%

words related to winning and success

No Comments

No Known Activations