INDEX

Explanations

terms related to negative or unwanted events or situations

oai_token-act-pair · gpt-3.5-turbo

terms related to unwanted or unwelcome situations and actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

alach

-0.90

bleacher

-0.79

hetti

-0.79

clerosis

-0.78

utenant

-0.77

utra

-0.77

sis

-0.77

lass

-0.76

bard

-0.76

ramid

-0.76

POSITIVE LOGITS

 pregnancies

1.16

 unwanted

1.04

 pregnancy

0.98

 Parenthood

0.89

ãĤī

0.89

 adolesc

0.82

 offspring

0.77

 unintended

0.75

 consequence

0.74

ãģĭ

0.73

Activations Density 0.015%

terms related to negative or unwanted events or situations

terms related to unwanted or unwelcome situations and actions

No Comments

No Known Activations