INDEX

Explanations

instances where attention is being directed or should be directed towards something

oai_token-act-pair · gpt-3.5-turbo

instances of paying attention in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

venge

-0.76

geries

-0.71

rie

-0.70

byn

-0.68

iHUD

-0.68

joining

-0.65

anus

-0.64

rafted

-0.63

headers

-0.62

cluding

-0.62

POSITIVE LOGITS

 attent

0.86

é¾įå¥ĳå£«

0.84

ibly

0.73

fulness

0.72

geist

0.68

 cues

0.67

IBLE

0.65

 attentive

0.64

 Listen

0.63

 quizz

0.63

Activations Density 0.020%

instances where attention is being directed or should be directed towards something

instances of paying attention in various contexts

No Comments

No Known Activations