INDEX

Explanations

words related to potential risks, dangers, challenges, and issues

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

words that describe threats or negative impacts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

swick

-0.83

mingham

-0.74

vez

-0.73

iban

-0.73

Cola

-0.69

livious

-0.68

irrel

-0.67

Nation

-0.66

vind

-0.65

flush

-0.65

POSITIVE LOGITS

 posed

1.17

 inherent

1.16

 associated

1.14

 caused

1.11

 inflicted

1.05

 endured

1.04

 attendant

1.03

 faced

1.02

 incurred

1.02

 encountered

1.01

Activations Density 0.238%

words related to potential risks, dangers, challenges, and issues

words that describe threats or negative impacts

No Comments

No Known Activations