INDEX

Explanations

proper nouns related to individuals or organizations

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names that appear frequently in the text

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

orie

-0.92

eric

-0.85

oral

-0.84

othe

-0.73

aned

-0.73

ocrats

-0.72

atics

-0.71

inx

-0.70

othes

-0.70

ane

-0.69

POSITIVE LOGITS

gerald

1.01

lein

0.85

 Braun

0.81

heimer

0.78

 Bunny

0.74

patrick

0.73

wald

0.73

ufact

0.73

hler

0.71

minster

0.71

Activations Density 0.024%

proper nouns related to individuals or organizations

proper nouns, particularly names that appear frequently in the text

No Comments

No Known Activations