INDEX

Explanations

words related to problems, challenges, or negative aspects

oai_token-act-pair · gpt-3.5-turbo

negative or problematic concepts and outcomes

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

NESS

-0.69

 Mellon

-0.68

WARD

-0.68

 Guilty

-0.67

into

-0.63

ILL

-0.62

Ħ¢

-0.61

late

-0.60

]"

-0.60

hower

-0.60

POSITIVE LOGITS

been

0.90

 knack

0.81

 chance

0.77

icum

0.73

kees

0.68

 opportunity

0.67

 flexibility

0.67

urred

0.65

 privileges

0.65

 grasp

0.64

Activations Density 0.433%

words related to problems, challenges, or negative aspects

negative or problematic concepts and outcomes

No Known Activations