INDEX

Explanations

references to prominent and influential figures

oai_token-act-pair · gpt-3.5-turbo

the word "pin" 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

anwhile

-0.76

 Norn

-0.72

mble

-0.67

mberg

-0.67

LIA

-0.67

 SERV

-0.67

Ö¼

-0.66

HUM

-0.65

 HELP

-0.63

Cf

-0.63

POSITIVE LOGITS

ning

1.22

pin

1.06

ned

1.05

ners

0.93

pins

0.92

folio

0.85

pole

0.84

wheel

0.83

ochet

0.83

cy

0.80

Activations Density 0.005%

references to prominent and influential figures

the word "pin" in various contexts

No Comments

No Known Activations