INDEX

Explanations

phrases indicating prohibition or negation

oai_token-act-pair · gpt-3.5-turbo

negations about what should or shouldn't occur

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

FW

-0.66

Via

-0.65

 Hungry

-0.64

Wid

-0.64

 Panasonic

-0.63

hole

-0.62

 irresistible

-0.61

LIN

-0.60

 Planes

-0.60

 marvelous

-0.59

POSITIVE LOGITS

 emulate

0.88

be

0.82

regnancy

0.78

reated

0.77

imize

0.76

 imitate

0.73

arent

0.73

erest

0.72

 necessarily

0.69

 behave

0.69

Activations Density 0.090%

phrases indicating prohibition or negation

negations about what should or shouldn't occur

No Comments

No Known Activations