INDEX

Explanations

phrases related to a specific purpose or goal

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing conclusions or implications in discussions

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

 vulner

-0.67

 Shares

-0.61

ãĤ¹ãĥĪ

-0.59

Osw

-0.59

azard

-0.56

 Dragons

-0.56

 TAMADRA

-0.56

 Tsukuyomi

-0.55

Dor

-0.53

 bets

-0.53

POSITIVE LOGITS

 extent

1.09

venge

0.87

 detriment

0.85

Ĥİ

0.77

rouse

0.75

merce

0.75

£ı

0.69

ilts

0.67

ffer

0.66

 tune

0.65

Activations Density 0.182%

phrases related to a specific purpose or goal

phrases emphasizing conclusions or implications in discussions

No Comments

No Known Activations