INDEX

Explanations

Easy-to-use technology-related terms

oai_token-act-pair · gpt-3.5-turbo

phrases related to ease of use

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

hips

-0.84

eters

-0.83

reon

-0.80

mbuds

-0.71

grave

-0.70

arians

-0.69

yss

-0.67

geist

-0.67

akin

-0.65

 Tyrann

-0.64

POSITIVE LOGITS

Jet

0.87

going

0.84

wallet

0.76

 prey

0.74

ãĥīãĥ©

0.74

ily

0.72

 minded

0.69

 easy

0.69

coded

0.68

 accessible

0.67

Activations Density 0.033%

Easy-to-use technology-related terms

phrases related to ease of use

No Comments

No Known Activations