INDEX

Explanations

words related to disguise or hiding

oai_token-act-pair · gpt-3.5-turbo

the presence of the word "gu"

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

hower

-0.93

ŃĶ

-0.80

 Spectrum

-0.77

 spect

-0.70

cling

-0.67

 eleph

-0.67

 Parables

-0.65

 shutter

-0.65

TTL

-0.63

baugh

-0.62

POSITIVE LOGITS

arant

1.12

idelines

1.06

arding

1.04

ilty

1.01

vernment

0.98

ests

0.98

arded

0.96

cci

0.96

inea

0.95

errilla

0.95

Activations Density 0.010%

words related to disguise or hiding

the presence of the word "gu"

No Comments

No Known Activations