INDEX

Explanations

mentions of formal approval or authorization

oai_token-act-pair · gpt-3.5-turbo

references to approval ratings or processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

quer

-0.67

rete

-0.67

pots

-0.67

»Ĵ

-0.66

Pie

-0.66

slice

-0.65

wic

-0.65

range

-0.65

ram

-0.64

adows

-0.63

POSITIVE LOGITS

 approval

3.89

 approvals

2.85

 disapproval

2.16

 approve

2.11

 approving

2.11

 Appro

1.97

 authorization

1.88

 permission

1.85

 approved

1.84

 approves

1.81

Activations Density 0.014%

mentions of formal approval or authorization

references to approval ratings or processes

No Comments

No Known Activations