INDEX

Explanations

words related to legal or governmental actions, specifically involving signatures and efforts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "to" indicating intent or purpose

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

yip

-0.90

 tert

-0.68

psey

-0.66

VIDIA

-0.64

ophon

-0.64

iosyncr

-0.62

MU

-0.60

RO

-0.60

OPER

-0.60

ambers

-0.60

POSITIVE LOGITS

Constructed

0.74

 Coat

0.66

jad

0.66

hur

0.65

bold

0.64

 Telegram

0.64

eded

0.63

elf

0.62

quartered

0.62

eting

0.61

Activations Density 0.000%

words related to legal or governmental actions, specifically involving signatures and efforts

instances of the word "to" indicating intent or purpose

No Comments

No Known Activations