INDEX

Explanations

mathematical expressions and code snippets that are formatted with specific characters

oai_token-act-pair · gpt-3.5-turbo

symbols and characters often used in coding or programming contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

teenth

-0.81

orian

-0.80

etheless

-0.80

onial

-0.80

odium

-0.80

İĭ

-0.76

entimes

-0.75

yssey

-0.75

erville

-0.74

eca

-0.72

POSITIVE LOGITS

_>

1.56

=>

0.90

0.89

.<

0.87

Suc

0.74

(<

0.73

PsyNetMessage

0.70

 Preferences

0.69

 About

0.67

0.66

Activations Density 0.013%

mathematical expressions and code snippets that are formatted with specific characters

symbols and characters often used in coding or programming contexts

No Comments

No Known Activations