INDEX

Explanations

words related to unhealthy, processed, or low-quality food items

oai_token-act-pair · gpt-3.5-turbo

terms related to junk food

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Æ

-0.88

cing

-0.82

gypt

-0.68

oral

-0.68

BILITY

-0.67

ndum

-0.66

È

-0.65

IRO

-0.65

 Pesh

-0.64

 Lauder

-0.64

POSITIVE LOGITS

 junk

1.08

 heap

0.95

 dumps

0.89

 drawer

0.84

 pile

0.82

unks

0.80

ies

0.79

 mail

0.78

adelphia

0.78

etsu

0.77

Activations Density 0.007%

words related to unhealthy, processed, or low-quality food items

terms related to junk food

No Comments

No Known Activations