INDEX

Explanations

references to the original version of something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "original."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Â·Â·

-0.65

ebin

-0.65

doi

-0.64

leground

-0.64

trap

-0.63

watching

-0.62

lies

-0.62

heid

-0.62

zone

-0.62

ries

-0.61

POSITIVE LOGITS

 original

3.47

 originals

2.71

original

2.51

Original

2.23

 Original

2.21

 ORIG

1.67

 originally

1.55

 initial

1.48

 previous

1.27

 actual

1.21

Activations Density 0.020%

references to the original version of something

occurrences of the word "original."

No Comments

No Known Activations