INDEX

Explanations

URLs or references to URLs

oai_token-act-pair · gpt-3.5-turbo

mentions of URLs and related web addresses

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

ynski

-0.88

manship

-0.81

cffff

-0.79

rentice

-0.72

rament

-0.72

SPONSORED

-0.70

hma

-0.70

ctuary

-0.69

itary

-0.69

wagen

-0.68

POSITIVE LOGITS

URL

1.08

URI

1.02

 URLs

1.02

URL

0.90

url

0.85

 encoded

0.82

URI

0.78

Url

0.77

 prefix

0.75

 encoding

0.73

Activations Density 0.013%

URLs or references to URLs

mentions of URLs and related web addresses

No Comments

No Known Activations