INDEX

Explanations

suggestions or recommendations

oai_token-act-pair · gpt-3.5-turbo

suggestions or recommendations in the text

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

agos

-0.74

 Bengal

-0.68

ELD

-0.66

Ern

-0.65

TED

-0.65

anders

-0.64

 Notting

-0.64

 Sabha

-0.63

OSH

-0.62

Leod

-0.61

POSITIVE LOGITS

eele

0.83

ezvous

0.80

ħĭ

0.80

rompt

0.77

bably

0.73

edi

0.70

yip

0.69

 reconsider

0.69

awaru

0.68

 intervention

0.66

Activations Density 0.219%

suggestions or recommendations

suggestions or recommendations in the text

No Comments

No Known Activations