INDEX

Explanations

phrases indicating a need for a specific type of item or concept

oai_token-act-pair · gpt-3.5-turbo

phrases indicating types or categories of things

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

©¶æ

-0.64

 Bots

-0.63

 Niet

-0.61

aspers

-0.58

 Uganda

-0.57

 Eisen

-0.57

 Leone

-0.56

 Tackle

-0.56

Wyn

-0.56

ften

-0.56

POSITIVE LOGITS

of

0.98

luster

0.77

thing

0.75

itional

0.72

 thereof

0.71

Of

0.71

Of

0.69

THING

0.68

hearted

0.68

itious

0.67

Activations Density 0.047%

phrases indicating a need for a specific type of item or concept

phrases indicating types or categories of things

No Comments

No Known Activations