INDEX

Explanations

names or parts of names of individuals

oai_token-act-pair · gpt-3.5-turbo

instances of specific letter combinations or sequences

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

vp

-0.64

apor

-0.63

pta

-0.60

ATURE

-0.59

OIL

-0.58

 dunno

-0.56

 actionGroup

-0.56

llah

-0.55

eled

-0.55

 MacArthur

-0.55

POSITIVE LOGITS

velt

0.97

merce

0.82

lyn

0.74

Thumbnail

0.74

imaru

0.74

 trout

0.73

deals

0.71

estone

0.71

amia

0.69

kins

0.65

Activations Density 0.137%

names or parts of names of individuals

instances of specific letter combinations or sequences

No Comments

No Known Activations