INDEX

Explanations

phrases related to opinions or positions on certain issues

oai_token-act-pair · gpt-3.5-turbo

expressions of opinions or positions on various subjects

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

NetMessage

-0.85

duct

-0.69

cing

-0.68

 Tycoon

-0.68

STON

-0.67

athan

-0.65

esters

-0.64

rafted

-0.63

amn

-0.63

IVERS

-0.63

POSITIVE LOGITS

 stance

1.30

 stances

1.16

 toward

0.92

 posture

0.86

 towards

0.86

 positions

0.85

 position

0.85

 reversal

0.78

 views

0.75

 pledge

0.74

Activations Density 0.024%

phrases related to opinions or positions on certain issues

expressions of opinions or positions on various subjects

No Comments

No Known Activations