INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

the name "Stephen."

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

 batches

-0.72

duino

-0.71

uez

-0.66

 offsets

-0.65

 dams

-0.64

 entrusted

-0.64

pil

-0.63

 buses

-0.63

 Mata

-0.63

 retrieval

-0.62

POSITIVE LOGITS

 Stephen

3.35

Stephen

2.87

 Steph

1.64

 Steve

1.62

 Steven

1.61

 Justin

1.53

 Stephenson

1.52

 Kevin

1.46

Steve

1.40

 Margaret

1.37

Activations Density 0.013%

proper nouns

the name "Stephen."

No Comments

No Known Activations