INDEX

Explanations

phrases related to consuming drinks and apparel

oai_token-act-pair · gpt-3.5-turbo

references to types of beverages, especially alcoholic ones

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

urion

-0.87

heses

-0.83

boxing

-0.82

urers

-0.80

bryce

-0.78

Reviewer

-0.78

projects

-0.76

brates

-0.74

urches

-0.72

hran

-0.72

POSITIVE LOGITS

 wine

1.05

 vodka

1.04

 soda

1.03

 beer

1.00

 champagne

0.99

 whiskey

0.98

 milk

0.95

 coffee

0.92

tea

0.91

 liquor

0.91

Activations Density 0.091%

phrases related to consuming drinks and apparel

references to types of beverages, especially alcoholic ones

No Comments

No Known Activations