INDEX

Explanations

words related to technological tools or software programs

oai_token-act-pair · gpt-3.5-turbo

placeholders or empty strings in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

bledon

-0.69

 Seym

-0.67

Ire

-0.66

Eth

-0.65

Fax

-0.63

erenn

-0.62

ç«

-0.61

Desk

-0.60

 Frie

-0.57

 Borders

-0.57

POSITIVE LOGITS

esson

0.74

 includes

0.73

 consists

0.72

 consisted

0.68

 violates

0.67

 utilizes

0.67

 lasted

0.66

 involves

0.65

 awaits

0.64

cius

0.63

Activations Density 0.055%

words related to technological tools or software programs

placeholders or empty strings in the text

No Comments

No Known Activations