INDEX

Explanations

relevant information by searching and finding specific terms or data in various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of discovery or retrieval of information

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 salute

-0.78

ISION

-0.75

stroke

-0.69

 halftime

-0.67

Exit

-0.65

Chief

-0.64

inth

-0.63

ornings

-0.62

 disrupting

-0.59

 disrupt

-0.59

POSITIVE LOGITS

 Wikipedia

1.49

wik

1.41

 youtube

1.41

 websites

1.38

 google

1.33

 blogs

1.32

 tutorials

1.30

 forums

1.28

 datas

1.28

 Youtube

1.28

Activations Density 0.494%

relevant information by searching and finding specific terms or data in various contexts

instances of discovery or retrieval of information

No Comments

No Known Activations