INDEX

Explanations

phrases with hyphenated words indicating a connection between different concepts

oai_token-act-pair · gpt-3.5-turbo

prefixes indicating negative or diminished conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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¥µ

-0.70

 McKenna

-0.67

è£ħ

-0.67

 Kali

-0.66

è¦ļéĨĴ

-0.65

 belts

-0.65

 Mock

-0.65

 Immunity

-0.64

 Kills

-0.64

 Warriors

-0.63

POSITIVE LOGITS

ruction

1.04

ipment

1.01

anced

0.99

anc

0.94

yright

0.93

tenance

0.93

ake

0.93

ousand

0.90

otic

0.89

artment

0.89

Activations Density 0.154%

phrases with hyphenated words indicating a connection between different concepts

prefixes indicating negative or diminished conditions

No Comments

No Known Activations