INDEX

Explanations

names, potentially with some focus on surnames

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names and entities

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

Not in Any Lists

No Comments

Negative Logits

stakes

-0.68

âĶĢâĶĢ

-0.64

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.64

 Tacoma

-0.64

 Mayhem

-0.63

Dee

-0.61

 Stafford

-0.59

WARE

-0.59

 Printed

-0.58

 Mankind

-0.58

POSITIVE LOGITS

auld

1.08

ica

0.97

hem

0.95

ault

0.91

aceae

0.91

akh

0.88

Arn

0.88

aldo

0.87

canon

0.86

issan

0.86

Activations Density 0.009%

names, potentially with some focus on surnames

mentions of specific names and entities

No Comments

No Known Activations