INDEX

Explanations

mentions of specific clothing items, particularly dresses

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "dress."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ntil

-0.82

 Blessed

-0.78

JV

-0.65

ocalyptic

-0.62

interrupted

-0.61

 Cursed

-0.61

rehend

-0.61

exist

-0.60

 condemned

-0.59

ventures

-0.58

POSITIVE LOGITS

maker

0.96

 dresses

0.96

 Dress

0.94

glers

0.93

 gown

0.92

 rehearsal

0.92

bag

0.90

 attire

0.89

 shoes

0.87

 dress

0.86

Activations Density 0.009%

mentions of specific clothing items, particularly dresses

repeated mentions of the word "dress."

No Comments

No Known Activations