INDEX

Explanations

words related to setting up, specifically focusing on the action of initiating or preparing something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "let" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

POSE

-0.78

ĻĤ

-0.78

 holiest

-0.72

PLIED

-0.72

antz

-0.71

natureconservancy

-0.71

 cumbers

-0.69

emetery

-0.67

 heavily

-0.62

 tremend

-0.62

POSITIVE LOGITS

let

1.02

ting

0.91

rack

0.81

istical

0.80

roll

0.79

cheon

0.78

erest

0.77

tering

0.76

imore

0.76

 lett

0.75

Activations Density 0.010%

words related to setting up, specifically focusing on the action of initiating or preparing something

instances of the word "let" in various contexts

No Comments

No Known Activations