INDEX

Explanations

phrases related to tasks and activities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating significant cultural or emotional impact

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

itton

-0.61

rather

-0.54

not

-0.51

 sandwic

-0.50

oso

-0.50

 proverb

-0.49

little

-0.47

 assorted

-0.46

JM

-0.46

almost

-0.44

POSITIVE LOGITS

 anymore

1.74

nor

1.68

 anything

1.16

any

1.13

 anywhere

1.08

 whatsoever

1.06

 slightest

1.02

 anybody

1.00

nor

0.96

ANY

0.94

Activations Density 1.509%

phrases related to tasks and activities

phrases indicating significant cultural or emotional impact

No Comments

No Known Activations