INDEX

Explanations

references to legal protections or rights

oai_token-act-pair · gpt-3.5-turbo

terms related to protections and rights in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hus

-0.73

thus

-0.67

 istg

-0.66

ker

-0.65

sis

-0.65

ergy

-0.64

leaf

-0.64

bus

-0.63

sb

-0.63

nexus

-0.62

POSITIVE LOGITS

 protections

1.46

 afforded

1.00

 safeguards

0.99

 protection

0.94

 Protect

0.90

 enshr

0.89

é¾įå¥ĳå£«

0.88

 protects

0.85

 disadvant

0.84

 protecting

0.84

Activations Density 0.014%

references to legal protections or rights

terms related to protections and rights in various contexts

No Comments

No Known Activations