INDEX

Explanations

the word "talented" with varying degrees of activation

oai_token-act-pair · gpt-3.5-turbo

the word "talented" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

FACE

-0.68

 Cheong

-0.67

ETA

-0.65

alone

-0.65

cart

-0.65

adish

-0.63

ocene

-0.63

classic

-0.62

intestinal

-0.60

ptions

-0.60

POSITIVE LOGITS

 talents

1.11

 talent

0.99

glers

0.90

 minds

0.89

anooga

0.83

nesses

0.80

 talented

0.80

ness

0.80

 entreprene

0.77

 minded

0.77

Activations Density 0.010%

the word "talented" with varying degrees of activation

the word "talented" in various contexts

No Comments

No Known Activations