INDEX

Explanations

words related to negative connotations or harmful attributes

oai_token-act-pair · gpt-3.5-turbo

terms associated with negative outcomes or harmful effects

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

 Carbuncle

-0.76

æĸ¹

-0.75

 Polo

-0.72

BOOK

-0.72

 Annotations

-0.72

ALK

-0.71

externalActionCode

-0.70

ãĥ¼ãĥĨãĤ£

-0.69

uyomi

-0.68

 Defenders

-0.67

POSITIVE LOGITS

colm

1.10

adies

1.10

ignant

1.06

formed

1.03

arial

0.96

igned

0.93

practice

0.93

ady

0.88

icious

0.86

mal

0.85

Activations Density 0.015%

words related to negative connotations or harmful attributes

terms associated with negative outcomes or harmful effects

No Comments

No Known Activations