INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

proper names and entities, particularly those related to people and organizations

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

âĶĢâĶĢ

-0.83

 deduction

-0.67

 Indo

-0.65

Â·Â·

-0.63

 underwater

-0.61

 archived

-0.60

;;;;;;;;;;;;

-0.59

 seriously

-0.59

 incub

-0.58

 Atom

-0.56

POSITIVE LOGITS

elman

1.12

ucci

1.00

angelo

1.00

igl

0.97

allo

0.96

ansky

0.96

illard

0.96

oley

0.96

aldi

0.96

acci

0.95

Activations Density 0.118%

proper nouns

proper names and entities, particularly those related to people and organizations

No Comments

No Known Activations