INDEX

Explanations

technical or detailed descriptions in content, such as discussing data, recipes, or instructions

oai_token-act-pair · gpt-3.5-turbo

expressions related to measurable success or outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Wellington

-0.60

 NASL

-0.57

 honoured

-0.55

 Fifa

-0.54

...

-0.52

 Pegasus

-0.52

 Hercules

-0.52

 Nasa

-0.51

 intestine

-0.51

 hemisphere

-0.51

POSITIVE LOGITS

âĢ

1.73

Â»

1.54

ãĢ

1.46

Â«

1.42

âĺ

1.38

âľ

1.37

Â¨

1.37

âĹı

1.36

</

1.36

âĶ

1.35

Activations Density 4.076%

technical or detailed descriptions in content, such as discussing data, recipes, or instructions

expressions related to measurable success or outcomes

No Comments

No Known Activations