INDEX

Explanations

the word "document."

oai_token-act-pair · gpt-3.5-turbo

references to various documents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

avorite

-0.87

olls

-0.83

cffff

-0.82

alsh

-0.73

 grav

-0.72

ntil

-0.70

ategory

-0.70

tones

-0.69

»Ĵ

-0.69

 actionGroup

-0.68

POSITIVE LOGITS

 document

1.18

arians

1.00

 documents

0.96

document

0.96

arian

0.95

 Document

0.93

ually

0.92

Document

0.84

ORS

0.83

DOC

0.78

Activations Density 0.011%

the word "document."

references to various documents

No Comments

No Known Activations