INDEX

Explanations

words related to design and architecture

oai_token-act-pair · gpt-3.5-turbo

references to design concepts and practices

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

ieri

-0.80

 Lauder

-0.78

nikov

-0.78

ICAN

-0.78

 Sheen

-0.74

essional

-0.74

 Witnesses

-0.74

Ago

-0.74

 Arist

-0.73

selves

-0.72

POSITIVE LOGITS

ators

1.04

yout

0.98

ator

0.97

ations

0.89

 design

0.86

ates

0.86

ating

0.86

 flaw

0.85

ated

0.84

 designs

0.80

Activations Density 0.035%

words related to design and architecture

references to design concepts and practices

No Comments

No Known Activations