INDEX

Explanations

specific patterns found in computer programming or code

oai_token-act-pair · gpt-3.5-turbo

ellipses or repeated punctuation 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

entimes

-0.94

 suspic

-0.84

 Lomb

-0.76

 clashed

-0.73

 manif

-0.70

 discredit

-0.69

 raged

-0.68

 traged

-0.68

 staggered

-0.68

 stagn

-0.67

POSITIVE LOGITS

 âĢİ

0.78

Done

0.74

BRE

0.71

 Bride

0.70

 âĢº

0.66

Hig

0.66

 Directive

0.65

 Appears

0.65

Daddy

0.65

rock

0.65

Activations Density 0.035%

specific patterns found in computer programming or code

ellipses or repeated punctuation in the text

No Comments

No Known Activations