INDEX

Explanations

phrases indicating a specific relationship or context between different entities

oai_token-act-pair · gpt-3.5-turbo

phrases concerning specific conditions or situations that start with "which."

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

bug

-0.67

Uk

-0.66

ax

-0.64

áµ

-0.62

Charg

-0.62

 Charg

-0.62

stage

-0.62

 Princ

-0.62

ben

-0.61

ma

-0.61

POSITIVE LOGITS

soever

0.85

xual

0.78

aceutical

0.75

velength

0.70

 dearly

0.69

earch

0.68

 Mercy

0.67

kson

0.67

 sake

0.66

edience

0.64

Activations Density 0.027%

phrases indicating a specific relationship or context between different entities

phrases concerning specific conditions or situations that start with "which."

No Comments

No Known Activations