INDEX

Explanations

references to various "Ve" trophies

oai_token-act-pair · gpt-3.5-turbo

references to the Vezina Trophy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

matically

-0.94

matic

-0.83

£ı

-0.82

manship

-0.80

hops

-0.78

eering

-0.77

 teasp

-0.71

ships

-0.68

itatively

-0.68

dress

-0.68

POSITIVE LOGITS

ggie

1.15

lez

1.12

Ve

1.06

Ve

1.00

ices

0.98

lette

0.95

endor

0.92

ggies

0.92

uve

0.91

Äį

0.90

Activations Density 0.010%

references to various "Ve" trophies

references to the Vezina Trophy

No Comments

No Known Activations