INDEX

Explanations

proper nouns related to locations or people

oai_token-act-pair · gpt-3.5-turbo

references to the word "Col" followed by a numeric designation, likely indicating a location or entity

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

cffffcc

-0.76

 Werewolf

-0.75

DragonMagazine

-0.74

 forth

-0.72

wered

-0.71

natureconservancy

-0.67

visor

-0.66

 Potential

-0.65

WHO

-0.65

iscopal

-0.64

POSITIVE LOGITS

ombo

1.04

ossal

0.98

Col

0.95

icol

0.94

onial

0.91

onel

0.91

leen

0.90

umbo

0.87

ophon

0.84

ado

0.82

Activations Density 0.003%

proper nouns related to locations or people

references to the word "Col" followed by a numeric designation, likely indicating a location or entity

No Comments

No Known Activations