INDEX

Explanations

instances where something is described or presented in a particular way

oai_token-act-pair · gpt-3.5-turbo

phrases related to the concept of framing or perspective in discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

onge

-0.92

andowski

-0.77

withstanding

-0.76

ibaba

-0.74

issy

-0.73

profit

-0.71

ggies

-0.71

ergic

-0.71

Mos

-0.71

clerosis

-0.70

POSITIVE LOGITS

eering

1.08

 framed

0.99

 framing

0.82

 frames

0.79

eers

0.75

Arg

0.71

 Frames

0.70

 Veil

0.70

 pane

0.69

 Vortex

0.68

Activations Density 0.010%

instances where something is described or presented in a particular way

phrases related to the concept of framing or perspective in discussions

No Comments

No Known Activations