INDEX

Explanations

mentions of the word "Ten"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Ten."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ready

-0.74

cffffcc

-0.72

 mileage

-0.64

userc

-0.64

****************

-0.63

lycer

-0.63

unct

-0.62

 mistaken

-0.61

 slic

-0.59

Ø©

-0.59

POSITIVE LOGITS

zin

1.09

acity

1.09

essee

0.98

acious

0.95

 Thousand

0.88

aciously

0.86

eta

0.83

ney

0.82

uta

0.78

ryu

0.78

Activations Density 0.011%

mentions of the word "Ten"

instances of the word "Ten."

No Comments

No Known Activations