INDEX

Explanations

entities or individuals that are well-known and respected in their respective fields

oai_token-act-pair · gpt-3.5-turbo

terms related to fame and recognition

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

adra

-0.86

plant

-0.83

anuts

-0.76

arettes

-0.75

eper

-0.75

cise

-0.75

 compuls

-0.74

heed

-0.74

plet

-0.73

ttle

-0.68

POSITIVE LOGITS

 renowned

0.93

 famed

0.90

 acclaimed

0.90

 BaseType

0.85

ESCO

0.83

 famous

0.70

 Dame

0.70

ieth

0.69

 renown

0.68

 legendary

0.68

Activations Density 0.013%

entities or individuals that are well-known and respected in their respective fields

terms related to fame and recognition

No Comments

No Known Activations