INDEX

Explanations

proper nouns, particularly names or titles

oai_token-act-pair · gpt-3.5-turbo

references to philosophical concepts or terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 footing

-0.61

jriwal

-0.61

 curtain

-0.61

 curtains

-0.61

ĸļ

-0.60

ciating

-0.60

eers

-0.59

 ledger

-0.58

 Documents

-0.58

 segregation

-0.58

POSITIVE LOGITS

adelphia

1.19

ebus

0.96

Ļ

0.87

ogram

0.80

aceutical

0.79

İ

0.74

¬

0.74

iston

0.74

agall

0.73

ashtra

0.72

Activations Density 0.105%

proper nouns, particularly names or titles

references to philosophical concepts or terminology

No Comments

No Known Activations