INDEX

Explanations

comparative phrases using "so" to describe intensity or extent

oai_token-act-pair · gpt-3.5-turbo

the word "so" in various contexts indicating emphasis or causation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ropolitan

-0.66

 glances

-0.62

 acre

-0.58

wreck

-0.57

 Arch

-0.54

 grunt

-0.53

MLA

-0.53

{:

-0.52

 Dive

-0.52

 Northwest

-0.52

POSITIVE LOGITS

bered

1.19

oths

1.10

oooo

1.06

oooooooooooooooo

1.04

ooo

1.03

oooooooo

1.00

zin

0.96

apy

0.95

othes

0.94

othe

0.93

Activations Density 0.092%

comparative phrases using "so" to describe intensity or extent

the word "so" in various contexts indicating emphasis or causation

No Comments

No Known Activations