INDEX

Explanations

phrases related to threats and dangers

oai_token-act-pair · gpt-3.5-turbo

negative descriptors related to safety or risk

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

krit

-0.64

ija

-0.62

leneck

-0.61

inas

-0.60

ordon

-0.60

AFP

-0.58

odox

-0.57

lp

-0.56

etsk

-0.55

miah

-0.54

POSITIVE LOGITS

to

1.62

to

1.44

To

1.18

TO

1.17

To

1.15

 thereto

1.11

 unto

0.85

ta

0.83

TO

0.78

toc

0.56

Activations Density 0.974%

phrases related to threats and dangers

negative descriptors related to safety or risk

No Comments

No Known Activations