INDEX

Explanations

terms related to large amounts or sizes, such as "meg"

oai_token-act-pair · gpt-3.5-turbo

references to "megawatts" or variations of power measurement terms

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

ãģį

-0.81

 Sense

-0.70

bable

-0.65

smanship

-0.65

 Desire

-0.63

âķĲâķĲ

-0.63

 Paradox

-0.63

Ô

-0.63

xual

-0.63

 Continued

-0.62

POSITIVE LOGITS

abytes

1.24

abyte

1.17

apixel

1.00

awatts

0.98

atron

0.94

apolis

0.93

acity

0.92

aton

0.91

atis

0.87

ambo

0.84

Activations Density 0.011%

terms related to large amounts or sizes, such as "meg"

references to "megawatts" or variations of power measurement terms

No Comments

No Known Activations