INDEX

Explanations

adverbs ending in 'ly'

oai_token-act-pair · gpt-3.5-turbo

adverbs that describe the manner of actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ilater

-0.84

unta

-0.75

 bats

-0.70

 Pirates

-0.68

icket

-0.67

eers

-0.65

roit

-0.64

Wiz

-0.63

iens

-0.63

eele

-0.62

POSITIVE LOGITS

 speaking

0.91

opsis

0.75

 populated

0.74

Ĥİ

0.73

miah

0.73

 translated

0.68

 opin

0.68

entimes

0.66

 approximate

0.65

tics

0.65

Activations Density 0.031%

adverbs ending in 'ly'

adverbs that describe the manner of actions

No Comments

No Known Activations