INDEX

Explanations

details about food recipes and cooking instructions

oai_token-act-pair · gpt-3.5-turbo

food-related terms and ingredients

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

 seism

-0.89

 Gamergate

-0.80

 USAF

-0.79

lihood

-0.77

 REPORT

-0.76

Student

-0.76

 GEAR

-0.75

 Petition

-0.74

 SpaceX

-0.72

 Lockheed

-0.72

POSITIVE LOGITS

 sauces

1.21

 chees

1.11

 sauce

1.10

 flavors

1.09

isine

1.07

 flavours

1.03

 pasta

1.02

 tofu

1.00

 salad

1.00

Ingredients

1.00

Activations Density 0.324%

details about food recipes and cooking instructions

food-related terms and ingredients

No Comments

No Known Activations