INDEX

Explanations

phrases related to regular or common practices

oai_token-act-pair · gpt-3.5-turbo

references to regular or habitual actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Straw

-0.68

 wedge

-0.68

oak

-0.66

oras

-0.65

ippi

-0.62

hips

-0.61

 Hemp

-0.59

wit

-0.59

arger

-0.58

utm

-0.58

POSITIVE LOGITS

 routine

1.50

 routines

1.22

outine

0.97

 regimen

0.80

zed

0.76

oleon

0.74

dayName

0.70

 maintenance

0.70

ERN

0.69

ogue

0.69

Activations Density 0.005%

phrases related to regular or common practices

references to regular or habitual actions

No Comments

No Known Activations