INDEX

Explanations

questions or unknown information

oai_token-act-pair · gpt-3.5-turbo

the word "what" and its repeated usage in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

robe

-0.71

enburg

-0.69

aches

-0.69

ulic

-0.68

ubs

-0.67

por

-0.64

trop

-0.63

cean

-0.63

ped

-0.63

gar

-0.63

POSITIVE LOGITS

soever

1.16

 happened

1.13

 happens

1.11

 sorts

1.10

 happ

1.05

 kinds

1.03

 transpired

0.93

 else

0.86

 exactly

0.85

 constitutes

0.84

Activations Density 0.118%

questions or unknown information

the word "what" and its repeated usage in various contexts

No Comments

No Known Activations