INDEX

Explanations

dates related to historical events

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols

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

 lasting

-0.74

 safe

-0.73

uania

-0.71

 safer

-0.66

gow

-0.66

paw

-0.66

 safest

-0.65

 volcano

-0.64

 levers

-0.63

lyn

-0.62

POSITIVE LOGITS

ãĥª

0.95

 Mellon

0.87

Downloadha

0.74

inton

0.74

acha

0.71

ipher

0.71

 Nept

0.70

)]

0.70

Ru

0.66

MIT

0.66

Activations Density 0.000%

dates related to historical events

special characters or symbols

No Comments

No Known Activations