INDEX

Explanations

words related to destruction or removal

oai_token-act-pair · gpt-3.5-turbo

words associated with deprivation or lack

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

 Dragonbound

-0.83

nings

-0.80

ãĤ¤ãĥĪ

-0.80

FORE

-0.79

 Hole

-0.77

 Flavoring

-0.75

ãĥīãĥ©ãĤ´ãĥ³

-0.74

 Millennium

-0.73

WORK

-0.72

 Sandwich

-0.71

POSITIVE LOGITS

utations

1.12

raved

1.11

utation

1.05

reci

1.03

ository

1.01

rec

1.01

ravity

0.96

orters

0.96

ugal

0.94

onent

0.90

Activations Density 0.010%

words related to destruction or removal

words associated with deprivation or lack

No Comments

No Known Activations