INDEX

Explanations

words related to granting permission or consent

oai_token-act-pair · gpt-3.5-turbo

instances of the word "approval" 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hed

-0.83

gm

-0.77

hing

-0.71

lla

-0.70

hester

-0.67

grave

-0.66

enegger

-0.66

Cop

-0.64

tal

-0.64

ãĤ©

-0.64

POSITIVE LOGITS

 approval

1.07

 approvals

1.03

 ratings

1.01

 rating

0.97

 confir

0.89

 approves

0.89

worthy

0.86

 authorizing

0.85

 commissions

0.85

 unanimously

0.83

Activations Density 0.020%

words related to granting permission or consent

instances of the word "approval" in various contexts

No Comments

No Known Activations