INDEX

Explanations

words related to stealing or crime

oai_token-act-pair · gpt-3.5-turbo

words related to viewing or observing actions or events

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

 corpor

-0.66

cul

-0.66

ter

-0.62

 prof

-0.61

 mell

-0.60

 nicotine

-0.60

 Sour

-0.60

 Guinness

-0.59

 mort

-0.59

 neoc

-0.59

POSITIVE LOGITS

atching

4.01

atches

3.72

atched

3.10

ATCH

2.22

atch

2.19

atcher

1.95

ersion

1.11

aths

1.09

akes

1.02

ibling

1.00

Activations Density 0.014%

words related to stealing or crime

words related to viewing or observing actions or events

No Comments

No Known Activations