INDEX

Explanations

instances of the word "Helvetica."

oai_token-act-pair · gpt-3.5-turbo

references to the word "Hel" or its variations

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

HAEL

-0.77

Ħ¢

-0.68

æĦ

-0.67

rers

-0.66

EAR

-0.65

ALLY

-0.64

©¶æ¥µ

-0.64

 Cunning

-0.64

inction

-0.64

WD

-0.62

POSITIVE LOGITS

Hel

1.00

idon

0.99

mand

0.88

igon

0.84

iop

0.82

mut

0.82

ios

0.82

wings

0.81

wered

0.81

mes

0.80

Activations Density 0.005%

instances of the word "Helvetica."

references to the word "Hel" or its variations

No Comments

No Known Activations