INDEX

Explanations

specific numerical quantities or technical terms

oai_token-act-pair · gpt-3.5-turbo

occurrences of numerical data and quantities

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

MJ

-0.68

christ

-0.60

 resent

-0.59

llah

-0.59

zos

-0.59

 Stra

-0.58

Grab

-0.57

Hol

-0.57

\<

-0.56

aldehyde

-0.56

POSITIVE LOGITS

ãĥİ

0.72

 apiece

0.71

ãĥ¯ãĥ³

0.69

milo

0.67

 Difficulty

0.64

 IMAGES

0.62

 espresso

0.61

ĻĤ

0.61

incinn

0.60

idon

0.60

Activations Density 0.300%

specific numerical quantities or technical terms

occurrences of numerical data and quantities

No Comments

No Known Activations