INDEX

Explanations

ancient languages and names

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols

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

alez

-0.80

 Vald

-0.74

ovie

-0.68

 embargo

-0.67

 broker

-0.64

AGES

-0.64

ened

-0.63

 Clown

-0.63

 vans

-0.62

uador

-0.61

POSITIVE LOGITS

á¹

1.05

á¸

0.98

Äģ

0.97

ĩ

0.96

¹

0.92

Å«

0.92

Ä«

0.90

ternity

0.90

ĥ

0.89

Ê

0.85

Activations Density 0.009%

ancient languages and names

special characters or symbols

No Comments

No Known Activations