INDEX

Explanations

references to electronic circuits

oai_token-act-pair · gpt-3.5-turbo

mentions of "circuit."

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

minster

-0.81

ipple

-0.72

ACTED

-0.69

Hispanic

-0.67

ever

-0.65

 Haku

-0.64

ipeg

-0.64

ighter

-0.64

 IMAGES

-0.63

zl

-0.62

POSITIVE LOGITS

 breaker

1.07

 circuits

0.98

 circuit

0.94

 Circuit

0.86

eering

0.83

engers

0.79

 boards

0.78

cade

0.76

eer

0.75

 Racer

0.74

Activations Density 0.026%

references to electronic circuits

mentions of "circuit."

No Comments

No Known Activations