INDEX

Explanations

keywords related to meal times, specifically focused on lunch

oai_token-act-pair · gpt-3.5-turbo

mentions of lunch

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

ignty

-0.79

mental

-0.64

fitting

-0.64

oral

-0.64

 infertility

-0.63

 extrap

-0.63

Filename

-0.63

Downloadha

-0.63

guided

-0.62

Constructed

-0.62

POSITIVE LOGITS

eteria

0.97

bag

0.95

wich

0.93

bowl

0.93

 lunch

0.93

time

0.86

 Lunch

0.86

worms

0.84

 meal

0.84

aturday

0.82

Activations Density 0.010%

keywords related to meal times, specifically focused on lunch

mentions of lunch

No Comments

No Known Activations