INDEX

Explanations

the word "Miller" with different variations and contexts

oai_token-act-pair · gpt-3.5-turbo

the name "Miller."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

liest

-0.92

quo

-0.85

angular

-0.82

want

-0.76

initialized

-0.73

rous

-0.71

Brazil

-0.71

endant

-0.71

liness

-0.70

pired

-0.69

POSITIVE LOGITS

 Miller

1.06

 Lite

0.90

 Mayhem

0.83

hound

0.80

beer

0.74

endale

0.73

myra

0.73

quist

0.71

hawk

0.71

hurst

0.68

Activations Density 0.009%

the word "Miller" with different variations and contexts

the name "Miller."

No Comments

No Known Activations