INDEX

Explanations

words starting with "wo"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "wo" in various forms

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

Magikarp

-0.81

++++++++++++++++

-0.75

IUM

-0.75

oslov

-0.67

ividual

-0.67

 Horowitz

-0.67

âĸ¬

-0.65

ä»£

-0.65

idates

-0.65

itated

-0.64

POSITIVE LOGITS

efully

1.21

ofer

1.15

ollen

1.14

eful

1.14

ocom

1.01

olly

0.95

onder

0.91

asted

0.89

aken

0.88

jo

0.86

Activations Density 0.023%

words starting with "wo"

instances of the word "wo" in various forms

No Comments

No Known Activations