INDEX

Explanations

phrases where the speaker believes or is certain about something

oai_token-act-pair · gpt-3.5-turbo

references to knowledge or awareness

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

stellar

-0.74

 Friend

-0.64

 Monetary

-0.61

Member

-0.57

 Admir

-0.56

 Aden

-0.55

 Wheel

-0.54

represented

-0.54

hips

-0.53

credit

-0.53

POSITIVE LOGITS

alian

1.01

self

0.96

chy

0.96

unes

0.90

ain

0.89

's

0.87

 hurts

0.84

 happened

0.83

 exists

0.83

displayText

0.83

Activations Density 0.100%

phrases where the speaker believes or is certain about something

references to knowledge or awareness

No Comments

No Known Activations