INDEX

Explanations

prepositions indicating location or direction

oai_token-act-pair · gpt-3.5-turbo

phrases that express uncertainty or qualifications regarding statements

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

FTWARE

-0.68

Russ

-0.67

eria

-0.65

glass

-0.64

ships

-0.63

alpha

-0.63

Pac

-0.60

PAC

-0.60

quel

-0.59

gravity

-0.59

POSITIVE LOGITS

 least

1.34

onement

1.01

abase

0.89

 times

0.88

roph

0.88

ention

0.87

yp

0.86

 dusk

0.77

hens

0.77

 variance

0.76

Activations Density 0.241%

prepositions indicating location or direction

phrases that express uncertainty or qualifications regarding statements

No Comments

No Known Activations