INDEX

Explanations

phrases related to awards and recognitions

oai_token-act-pair · gpt-3.5-turbo

references to awards and recognitions, particularly those related to the year or time periods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agen

-0.71

aird

-0.71

enegger

-0.70

conn

-0.67

english

-0.67

nai

-0.66

APTER

-0.65

occupied

-0.64

duino

-0.64

actionDate

-0.63

POSITIVE LOGITS

 honors

1.21

 Award

1.18

 award

1.15

 contender

1.10

 awards

1.09

 accol

1.05

 nominee

1.02

 Awards

0.99

 nominees

0.98

 finalists

0.98

Activations Density 0.052%

phrases related to awards and recognitions

references to awards and recognitions, particularly those related to the year or time periods

No Comments

No Known Activations