INDEX

Explanations

references to people or things that are liked or favored

oai_token-act-pair · gpt-3.5-turbo

phrases that express preferences or things that people enjoy

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

INA

-0.81

ARM

-0.71

inas

-0.67

Decl

-0.67

Flight

-0.67

nex

-0.66

 Peninsula

-0.66

Frag

-0.66

arta

-0.66

edia

-0.65

POSITIVE LOGITS

lihood

1.48

 minded

0.98

 likes

0.89

ĸļ

0.87

terday

0.85

ngth

0.81

minded

0.80

ileaks

0.79

 liking

0.77

Activations Density 0.006%

references to people or things that are liked or favored

phrases that express preferences or things that people enjoy

No Comments

No Known Activations