INDEX

Explanations

instances of emphasis indicating a large quantity or degree

oai_token-act-pair · gpt-3.5-turbo

expressions emphasizing quantity or degree

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

İĭ

-0.85

etts

-0.84

ĺħ

-0.80

grad

-0.76

ologies

-0.73

jac

-0.70

oway

-0.70

bley

-0.69

Ń·

-0.68

sov

-0.67

POSITIVE LOGITS

ado

0.76

 effort

0.73

 appreciated

0.71

 attention

0.69

antz

0.66

 interest

0.64

 faith

0.64

 integrity

0.63

EAR

0.63

 conscience

0.62

Activations Density 0.022%

instances of emphasis indicating a large quantity or degree

expressions emphasizing quantity or degree

No Comments

No Known Activations