INDEX

Explanations

different categories or types of items within various contexts, such as risk, words, or services

oai_token-act-pair · gpt-3.5-turbo

phrases that categorize or describe items into groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

bats

-0.78

 Mechdragon

-0.69

uddin

-0.66

ãĤ¼ãĤ¦ãĤ¹

-0.63

liga

-0.62

 Annotations

-0.61

 Boxing

-0.61

Alz

-0.59

apeake

-0.59

 Archangel

-0.59

POSITIVE LOGITS

 distinct

0.78

%:

0.77

 simultaneously

0.76

viz

0.75

 sexes

0.73

rencies

0.72

:-

0.71

 sets

0.71

 thirds

0.71

:(

0.70

Activations Density 0.237%

different categories or types of items within various contexts, such as risk, words, or services

phrases that categorize or describe items into groups

No Comments

No Known Activations