INDEX

Explanations

references to the original publication or appearance of articles

oai_token-act-pair · gpt-3.5-turbo

references to article publication and authorship

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

knit

-0.74

 levers

-0.67

ibles

-0.67

gui

-0.65

 orderly

-0.65

 bystanders

-0.64

 trou

-0.63

$$$$

-0.62

bral

-0.61

 connectors

-0.60

POSITIVE LOGITS

 reprinted

0.91

Published

0.79

posted

0.78

 Publication

0.75

////

0.74

 Published

0.73

 Originally

0.72

Posted

0.71

 Download

0.69

Copyright

0.69

Activations Density 0.101%

references to the original publication or appearance of articles

references to article publication and authorship

No Comments

No Known Activations