INDEX

Explanations

words related to broad permissions or options, especially the word "any"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "any" and its variations in usage contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

vier

-0.77

illard

-0.68

omsday

-0.65

nas

-0.64

gers

-0.63

bows

-0.60

Leaks

-0.59

forth

-0.59

iatus

-0.59

 alot

-0.59

POSITIVE LOGITS

THING

1.31

WHERE

1.12

ONE

1.10

 kind

1.08

where

1.07

ones

1.01

body

1.00

 conceivable

1.00

 combination

0.99

 number

0.94

Activations Density 0.075%

words related to broad permissions or options, especially the word "any"

occurrences of the word "any" and its variations in usage contexts

No Comments

No Known Activations