INDEX

Explanations

mentions of official documents or investigations

oai_token-act-pair · gpt-3.5-turbo

words associated with formal documentation or processes

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

apo

-0.83

pan

-0.77

assies

-0.74

coat

-0.73

retch

-0.71

pled

-0.69

angu

-0.69

pert

-0.68

hov

-0.67

asive

-0.66

POSITIVE LOGITS

BY

0.81

 jointly

0.79

by

0.76

ãĤ´ãĥ³

0.74

 aback

0.71

hran

0.71

terness

0.70

 Called

0.65

tesy

0.65

 anonymously

0.64

Activations Density 0.300%

mentions of official documents or investigations

words associated with formal documentation or processes

No Comments

No Known Activations