INDEX

Explanations

the word "Reason" and related forms

oai_token-act-pair · gpt-3.5-turbo

instances and variations of the word "reason"

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

avorite

-0.82

ibaba

-0.79

omez

-0.72

 Carbuncle

-0.72

yss

-0.68

wana

-0.67

annis

-0.65

ModLoader

-0.63

 Watt

-0.62

 Wink

-0.62

POSITIVE LOGITS

abl

1.16

why

0.94

ably

0.88

boards

0.78

neum

0.77

Ľ

0.74

orial

0.73

¿½

0.73

WHY

0.71

atic

0.71

Activations Density 0.025%

the word "Reason" and related forms

instances and variations of the word "reason"

No Comments

No Known Activations