INDEX

Explanations

references to historical events or places

oai_token-act-pair · gpt-3.5-turbo

references to historical significance or events

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Coil

-0.82

worm

-0.76

Feed

-0.74

azor

-0.74

Qi

-0.74

zzle

-0.73

Switch

-0.72

puff

-0.71

worms

-0.68

 Anon

-0.67

POSITIVE LOGITS

 historic

3.34

historic

2.44

 Historic

2.34

 historical

2.10

 landmark

1.79

 historically

1.63

 Historical

1.63

 monumental

1.31

 watershed

1.30

 iconic

1.27

Activations Density 0.017%

references to historical events or places

references to historical significance or events

No Comments

No Known Activations