INDEX

Explanations

names or terms containing "dn."

oai_token-act-pair · gpt-3.5-turbo

terms that signify certainty or definitive knowledge

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

atories

-0.77

reth

-0.75

aud

-0.73

âĹ¼

-0.71

otaur

-0.69

eki

-0.67

atin

-0.65

iqu

-0.65

Ĩ

-0.64

inson

-0.64

POSITIVE LOGITS

etheless

0.96

ulhu

0.86

vernment

0.76

ignty

0.75

aido

0.73

urches

0.73

urchase

0.72

 Copenhagen

0.71

ipeg

0.70

 Inqu

0.69

Activations Density 0.033%

names or terms containing "dn."

terms that signify certainty or definitive knowledge

No Comments

No Known Activations