INDEX

Explanations

instances where the word "lid" is mentioned

oai_token-act-pair · gpt-3.5-turbo

references to lids or covers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

bec

-0.70

owitz

-0.66

onies

-0.66

 Seeking

-0.63

yan

-0.62

be

-0.61

plays

-0.61

sports

-0.60

assault

-0.60

arry

-0.60

POSITIVE LOGITS

lid

4.22

 curtain

1.18

 eyel

1.04

 valve

1.02

 latch

1.02

 veil

1.01

rim

0.99

 hinge

0.98

 nozzle

0.98

 seal

0.97

Activations Density 0.015%

instances where the word "lid" is mentioned

references to lids or covers

No Comments

No Known Activations