INDEX

Explanations

proper nouns denoting persons and their actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "left"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

wcsstore

-0.98

MAT

-0.84

è¦ļéĨĴ

-0.69

alist

-0.68

intend

-0.67

MpServer

-0.65

Percent

-0.64

inos

-0.63

agers

-0.63

 mathemat

-0.63

POSITIVE LOGITS

overs

0.98

 undone

0.89

 untreated

0.82

 unexpl

0.81

wich

0.78

wing

0.76

 Dise

0.73

 Leaves

0.70

 Aleppo

0.70

illes

0.69

Activations Density 0.041%

proper nouns denoting persons and their actions

instances of the word "left"

No Comments

No Known Activations