INDEX

Explanations

mathematical or programming expressions and content within the text

oai_token-act-pair · gpt-3.5-turbo

elements related to programming constructs, variables, and functions

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

 Chattanooga

-0.75

 Geoffrey

-0.73

 Cape

-0.71

 Head

-0.71

Elk

-0.68

 Venture

-0.66

 Mash

-0.66

 Gent

-0.66

 seiz

-0.65

 Gorge

-0.65

POSITIVE LOGITS

1.65

(-

1.55

($

1.52

(_

1.52

((

1.47

(\

1.40

([

1.39

({

1.39

('

1.38

("

1.35

Activations Density 0.069%

mathematical or programming expressions and content within the text

elements related to programming constructs, variables, and functions

No Comments

No Known Activations