INDEX

Explanations

the word "conceivable" or variations of it

oai_token-act-pair · gpt-3.5-turbo

references to the concept of concealment or hiding information

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

 Bucks

-0.66

 GEAR

-0.65

KP

-0.65

 GOODMAN

-0.62

isms

-0.62

 Hour

-0.61

 waters

-0.59

OPS

-0.58

 Hayward

-0.58

ICS

-0.58

POSITIVE LOGITS

ivably

1.62

veland

1.05

iving

1.03

rence

1.01

aling

1.00

vable

1.00

avement

0.95

perate

0.92

abled

0.91

mble

0.90

Activations Density 0.037%

the word "conceivable" or variations of it

references to the concept of concealment or hiding information

No Comments

No Known Activations