INDEX

Explanations

mentions of the word "Char" within text

oai_token-act-pair · gpt-3.5-turbo

references to a specific character named "Char."

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

æĸ¹

-0.82

 Reloaded

-0.80

CLASSIFIED

-0.67

WARD

-0.66

 AMERICA

-0.64

ItemTracker

-0.63

 lodge

-0.62

 lockout

-0.61

EMS

-0.60

hower

-0.60

POSITIVE LOGITS

acters

1.31

itably

1.17

coal

1.11

isma

1.09

itable

1.06

iton

1.05

vin

0.92

izard

0.91

rier

0.91

iatric

0.90

Activations Density 0.022%

mentions of the word "Char" within text

references to a specific character named "Char."

No Comments

No Known Activations