INDEX

Explanations

phrases indicating belief or speculation

oai_token-act-pair · gpt-3.5-turbo

phrases involving beliefs or assumptions about individuals or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Perspective

-0.68

 transformations

-0.66

case

-0.66

 Ratings

-0.65

 Reconstruction

-0.62

comes

-0.62

finder

-0.61

pointers

-0.61

Valid

-0.61

ements

-0.61

POSITIVE LOGITS

be

0.98

 resemble

0.94

 incorporate

0.92

 indicate

0.91

 imitate

0.88

 belong

0.88

 imply

0.87

 perform

0.87

 differentiate

0.85

 adhere

0.85

Activations Density 0.100%

phrases indicating belief or speculation

phrases involving beliefs or assumptions about individuals or events

No Comments

No Known Activations