INDEX

Explanations

phrases related to verifying humanity and avoiding being identified as a robot

oai_token-act-pair · gpt-3.5-turbo

phrases related to user verification and authentication

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aimon

-0.81

enza

-0.79

Js

-0.79

 appell

-0.72

conom

-0.72

hawks

-0.71

sers

-0.65

çļ

-0.65

 Bord

-0.64

etts

-0.63

POSITIVE LOGITS

authorized

0.70

');

0.66

 sudo

0.62

 authorized

0.61

 HTML

0.61

 subscribing

0.61

 content

0.60

 embed

0.60

email

0.60

 password

0.60

Activations Density 0.028%

phrases related to verifying humanity and avoiding being identified as a robot

phrases related to user verification and authentication

No Comments

No Known Activations