INDEX

Explanations

references to the name "Walt" followed by a number indicating associated significance

oai_token-act-pair · gpt-3.5-turbo

variations of the word 'alt' indicating different contexts or subjects

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

#$

-0.60

NING

-0.60

INTER

-0.59

BLE

-0.58

ãĤ©

-0.58

 Occupations

-0.57

 hell

-0.56

 Pigs

-0.56

 sequ

-0.55

ystem

-0.55

POSITIVE LOGITS

imore

1.30

reatment

1.15

zman

1.13

ogether

1.12

agne

0.99

itude

0.97

imes

0.97

ierra

0.96

itudes

0.94

zer

0.93

Activations Density 0.028%

references to the name "Walt" followed by a number indicating associated significance

variations of the word 'alt' indicating different contexts or subjects

No Comments

No Known Activations