INDEX

Explanations

phrases emphasizing one particular thing or concept

oai_token-act-pair · gpt-3.5-turbo

repeated references to "one thing."

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

mson

-0.73

iolet

-0.71

illon

-0.70

ocks

-0.69

ocked

-0.68

illus

-0.68

inders

-0.66

avascript

-0.66

zanne

-0.65

ICO

-0.65

POSITIVE LOGITS

 thing

0.96

 hundred

0.95

 exception

0.92

 Hundred

0.82

 thousand

0.78

 sided

0.77

 eyed

0.73

guy

0.73

 remaining

0.72

 person

0.69

Activations Density 0.052%

phrases emphasizing one particular thing or concept

repeated references to "one thing."

No Comments

No Known Activations