INDEX

Explanations

numeric variables or code snippets

oai_token-act-pair · gpt-3.5-turbo

references to navigational or structural information, such as paths or menus

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

emouth

-0.80

estern

-0.78

 Seym

-0.76

ened

-0.75

 nodd

-0.72

mble

-0.71

enthal

-0.71

eki

-0.71

rious

-0.71

 takedown

-0.70

POSITIVE LOGITS

PsyNetMessage

0.81

Pg

0.80

flags

0.75

OOL

0.74

Constructed

0.73

advertising

0.71

=#

0.70

Available

0.68

 justify

0.67

--+

0.67

Activations Density 0.012%

numeric variables or code snippets

references to navigational or structural information, such as paths or menus

No Comments

No Known Activations