INDEX

Explanations

references to the foundational or fundamental aspects of something

oai_token-act-pair · gpt-3.5-turbo

phrases that discuss foundational arguments or justifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

opers

-0.87

oping

-0.79

ipop

-0.73

cha

-0.73

 Wend

-0.72

oho

-0.71

 Rosenberg

-0.71

estern

-0.69

andro

-0.68

osen

-0.68

POSITIVE LOGITS

 basis

0.93

 underpin

0.79

 foundation

0.79

plates

0.77

 grounding

0.75

 premise

0.75

 thereof

0.73

 grounds

0.72

 footing

0.71

 assumptions

0.69

Activations Density 0.031%

references to the foundational or fundamental aspects of something

phrases that discuss foundational arguments or justifications

No Comments

No Known Activations