INDEX

Explanations

questions ending with a question mark

oai_token-act-pair · gpt-3.5-turbo

questions aimed at engaging readers or prompting interaction

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

 scapego

-0.73

alty

-0.72

 reven

-0.71

aling

-0.71

als

-0.70

aper

-0.69

iculture

-0.69

riet

-0.68

 manif

-0.67

al

-0.66

POSITIVE LOGITS

 Nope

0.99

 Check

0.92

 Wouldn

0.87

Yep

0.85

 Maybe

0.85

 Probably

0.83

 Answer

0.82

 Think

0.82

 Well

0.82

 Sure

0.81

Activations Density 0.096%

questions ending with a question mark

questions aimed at engaging readers or prompting interaction

No Comments

No Known Activations