INDEX

Explanations

objects or locations related to furniture

oai_token-act-pair · gpt-3.5-turbo

instances of the word "table" and its context

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

vernment

-1.03

trak

-0.77

 Directorate

-0.73

ï¸ı

-0.70

 Archangel

-0.67

irtual

-0.66

é»Ĵ

-0.65

inventoryQuantity

-0.65

 Enhancement

-0.64

ust

-0.64

POSITIVE LOGITS

cloth

1.59

tops

1.07

table

0.98

 table

0.95

 tables

0.87

top

0.87

room

0.82

sheet

0.81

stairs

0.80

 stool

0.79

Activations Density 0.012%

objects or locations related to furniture

instances of the word "table" and its context

No Comments

No Known Activations