INDEX

Explanations

instances where something is being contrasted or compared with others

oai_token-act-pair · gpt-3.5-turbo

references to contrasting groups or opinions

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

-+

-0.72

 Purchase

-0.71

Accessory

-0.70

Deal

-0.67

isner

-0.66

opoly

-0.64

âĹ¼

-0.64

ihara

-0.63

Dub

-0.63

-0.60

POSITIVE LOGITS

ngth

0.88

etheless

0.88

ect

0.76

yout

0.72

rail

0.71

ierre

0.68

otom

0.65

sheets

0.64

pite

0.62

nep

0.62

Activations Density 0.089%

instances where something is being contrasted or compared with others

references to contrasting groups or opinions

No Comments

No Known Activations