INDEX

Explanations

words related to people's surnames

oai_token-act-pair · gpt-3.5-turbo

instances of the word "les" in various contexts

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

 isolation

-0.69

 attribution

-0.69

 isol

-0.68

 attribut

-0.67

raviolet

-0.64

 pund

-0.63

tains

-0.61

 vector

-0.61

 congress

-0.61

 everyday

-0.60

POSITIVE LOGITS

bury

1.15

heed

1.11

ttes

1.10

hots

1.10

bian

1.08

bians

1.06

mith

1.04

les

1.02

creen

1.00

hip

0.99

Activations Density 0.009%

words related to people's surnames

instances of the word "les" in various contexts

No Comments

No Known Activations