INDEX

Explanations

instances where something is shown or demonstrated

oai_token-act-pair · gpt-3.5-turbo

phrases indicating evidence or demonstrations of concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

icum

-0.79

osate

-0.76

rit

-0.74

cel

-0.73

assisted

-0.73

ilit

-0.73

lake

-0.72

orrow

-0.72

idis

-0.71

rongh

-0.70

POSITIVE LOGITS

 shown

1.04

¶æ

0.88

 prominently

0.88

ĸļ

0.81

=>

0.78

obin

0.76

 Thumbnails

0.75

shows

0.74

@#

0.73

Īè

0.73

Activations Density 0.010%

instances where something is shown or demonstrated

phrases indicating evidence or demonstrations of concepts

No Comments

No Known Activations