INDEX

Explanations

the letter "ñ" in text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the character "ñ" 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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 MISS

-0.78

ÃŁ

-0.75

lda

-0.71

leigh

-0.67

mble

-0.65

zzo

-0.62

HEAD

-0.61

SUP

-0.60

 eleph

-0.60

fusc

-0.60

POSITIVE LOGITS

Ã±

0.83

istry

0.74

Ã±a

0.74

erer

0.71

erville

0.70

Ã£

0.68

Ã³n

0.68

odied

0.68

rez

0.64

 Nieto

0.64

Activations Density 0.038%

the letter "ñ" in text

occurrences of the character "ñ" in various contexts

No Comments

No Known Activations