INDEX

Explanations

actions or situations of doing something without permission

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate actions or decisions taken without proper authorization or consent

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

è¦ļéĨĴ

-0.80

 Offline

-0.75

 Frag

-0.75

obyl

-0.72

 Reincarnated

-0.72

 fragmentation

-0.71

ãĤ¼ãĤ¦ãĤ¹

-0.70

 Falling

-0.67

aneers

-0.67

hov

-0.65

POSITIVE LOGITS

 permission

1.10

 consent

1.09

 lawful

1.08

 authorization

0.95

 provocation

0.94

 intending

0.92

 lawfully

0.90

 remorse

0.85

 intent

0.82

 NOTICE

0.82

Activations Density 0.412%

actions or situations of doing something without permission

phrases that indicate actions or decisions taken without proper authorization or consent

No Comments

No Known Activations