INDEX

Explanations

references to the word "Candle"

oai_token-act-pair · gpt-3.5-turbo

references to candles or related terms

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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.90

HF

-0.76

 pity

-0.69

ORGE

-0.67

 ENTER

-0.66

FactoryReloaded

-0.64

ãĤ´ãĥ³

-0.63

FG

-0.63

STD

-0.62

PUT

-0.60

POSITIVE LOGITS

idates

1.36

idate

1.10

lest

1.04

encies

0.89

acies

0.87

les

0.86

Cand

0.83

ice

0.82

illo

0.82

ulum

0.82

Activations Density 0.014%

references to the word "Candle"

references to candles or related terms

No Comments

No Known Activations