INDEX

Explanations

references to the word "original"

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "originality" or "original" content

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

wal

-0.75

walk

-0.74

rogens

-0.70

robe

-0.68

=-=-=-=-=-=-=-=-

-0.68

opping

-0.67

abad

-0.67

asia

-0.67

orney

-0.66

owler

-0.65

POSITIVE LOGITS

ITY

0.96

 incarnation

0.96

ity

0.84

 trilogy

0.83

 version

0.79

 batch

0.79

 impetus

0.78

ãĥīãĥ©ãĤ´ãĥ³

0.77

isations

0.75

istar

0.75

Activations Density 0.014%

references to the word "original"

references to the concept of "originality" or "original" content

No Comments

No Known Activations