INDEX

Explanations

the verb "do" with a high level of activation

oai_token-act-pair · gpt-3.5-turbo

instances of the verb "do."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Handling

-0.74

dayName

-0.74

Reviewer

-0.71

Software

-0.68

lights

-0.66

 Secondly

-0.64

Hok

-0.64

Tai

-0.62

thur

-0.62

 Quarter

-0.62

POSITIVE LOGITS

omsday

1.12

pez

1.07

ppel

1.01

lez

0.97

oley

0.89

ggy

0.86

omething

0.84

lyak

0.82

gging

0.82

oway

0.81

Activations Density 0.155%

the verb "do" with a high level of activation

instances of the verb "do."

No Comments

No Known Activations