INDEX

Explanations

the word "start" occurring in various contexts

oai_token-act-pair · gpt-3.5-turbo

phrases associated with starting or beginning events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

pedia

-0.70

illard

-0.68

ocally

-0.66

 majority

-0.63

phy

-0.63

asca

-0.62

inately

-0.61

selves

-0.59

 entirety

-0.58

 perfection

-0.57

POSITIVE LOGITS

nings

1.40

ners

0.91

ups

0.90

ribune

0.87

rek

0.84

points

0.83

up

0.82

UP

0.82

emaker

0.75

starting

0.73

Activations Density 0.043%

the word "start" occurring in various contexts

phrases associated with starting or beginning events

No Comments

No Known Activations