INDEX

Explanations

mentions of the abbreviation or location "WA."

oai_token-act-pair · gpt-3.5-turbo

references to the state of Washington

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

draw

-0.77

xual

-0.76

displayText

-0.74

ures

-0.72

lihood

-0.72

erous

-0.69

roman

-0.69

ãĤ¸

-0.69

 paths

-0.66

schild

-0.65

POSITIVE LOGITS

WA

1.09

WA

1.09

VE

0.94

HAHA

0.93

velength

0.90

ILA

0.85

TN

0.84

TX

0.83

LC

0.81

UGH

0.80

Activations Density 0.007%

mentions of the abbreviation or location "WA."

references to the state of Washington

No Comments

No Known Activations