INDEX

Explanations

keywords and phrases related to updates, versions, and fixes

oai_token-act-pair · gpt-3.5-turbo

titles or headings related to various topics in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 manif

-0.78

boa

-0.70

lene

-0.69

 acquaintance

-0.66

seys

-0.66

onite

-0.66

ĪĴ

-0.65

 subst

-0.64

 slump

-0.63

idine

-0.63

POSITIVE LOGITS

 WATCHED

0.86

 VIDEOS

0.83

 SHARES

0.79

 SPACE

0.78

 Expand

0.75

Why

0.75

 Cosponsors

0.75

 What

0.74

 Despite

0.73

 à¤

0.73

Activations Density 0.125%

keywords and phrases related to updates, versions, and fixes

titles or headings related to various topics in text

No Comments

No Known Activations