INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

names of characters and references to popular culture

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

itutional

-0.70

URA

-0.70

EMBER

-0.69

 elig

-0.69

IENT

-0.69

 tert

-0.68

 sympathetic

-0.65

 unde

-0.65

.''.

-0.65

 latitude

-0.64

POSITIVE LOGITS

 Soup

1.10

Girl

1.10

 Pants

1.09

Pie

1.07

Mania

1.03

 Alley

1.03

 Balls

1.02

 Bunny

1.02

pants

1.01

 Juice

1.00

Activations Density 0.313%

proper nouns

names of characters and references to popular culture

No Comments

No Known Activations