INDEX

Explanations

mentions of specific physical locations, particularly booths

oai_token-act-pair · gpt-3.5-turbo

references to booths or similar structures

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

ensions

-0.69

itals

-0.69

mitted

-0.68

ochond

-0.67

itars

-0.63

 acron

-0.62

ract

-0.61

ension

-0.60

mitting

-0.60

nesses

-0.59

POSITIVE LOGITS

 booth

1.08

 booths

0.97

holder

0.95

holders

0.92

cloth

0.87

washer

0.85

swer

0.80

handler

0.79

wagon

0.78

 exhib

0.76

Activations Density 0.013%

mentions of specific physical locations, particularly booths

references to booths or similar structures

No Comments

No Known Activations