INDEX

Explanations

names, likely related to credits or acknowledgements

oai_token-act-pair · gpt-3.5-turbo

instances of names or identifiers associated with individuals or groups in a specific context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 misunder

-0.71

cha

-0.71

otin

-0.70

interstitial

-0.69

zers

-0.68

casts

-0.67

cards

-0.66

planes

-0.65

pport

-0.65

Â¯

-0.64

POSITIVE LOGITS

Jr

1.13

MD

0.95

Sr

0.94

 Photographer

0.94

 Founder

0.93

 Editor

0.92

 Managing

0.91

PhD

0.90

 Owner

0.90

 Partner

0.90

Activations Density 0.117%

names, likely related to credits or acknowledgements

instances of names or identifiers associated with individuals or groups in a specific context

No Comments

No Known Activations