INDEX

Explanations

references to envelopes and related actions

oai_token-act-pair · gpt-3.5-turbo

mentions of envelopes and related concepts

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

anooga

-0.79

Ath

-0.71

ãĤ´ãĥ³

-0.71

onomous

-0.71

 Titans

-0.70

nesota

-0.69

 Nanto

-0.68

ç¥ŀ

-0.64

 Texans

-0.64

Dob

-0.62

POSITIVE LOGITS

 envelope

1.05

velop

1.03

mble

0.97

 envelop

0.93

eers

0.90

nuts

0.83

boxes

0.80

arding

0.78

 shaped

0.77

 mailbox

0.74

Activations Density 0.034%

references to envelopes and related actions

mentions of envelopes and related concepts

No Comments

No Known Activations