INDEX

Explanations

proper nouns related to personalities, especially those named Whitney and Whitman

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of specific names and terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

untarily

-0.85

icably

-0.78

MpServer

-0.72

auri

-0.71

 aver

-0.71

axis

-0.71

ogical

-0.68

ocre

-0.68

 fertil

-0.67

ugal

-0.67

POSITIVE LOGITS

 Whitney

1.29

 Whitman

1.01

Whit

0.93

jing

0.82

hua

0.77

mast

0.77

unciation

0.76

 Garage

0.76

iper

0.76

 ming

0.76

Activations Density 0.008%

proper nouns related to personalities, especially those named Whitney and Whitman

repeated mentions of specific names and terms

No Comments

No Known Activations