INDEX

Explanations

references to various names, likely looking for specific individuals or entities

oai_token-act-pair · gpt-3.5-turbo

prefixes and suffixes related to names and places

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

bley

-0.66

 goblin

-0.64

 stout

-0.59

 Dragonbound

-0.58

 swall

-0.57

cakes

-0.57

 horizont

-0.56

cake

-0.56

 braces

-0.55

 Sorceress

-0.55

POSITIVE LOGITS

ateg

0.78

rim

0.77

hea

0.77

az

0.73

aution

0.72

rit

0.71

ondo

0.71

ito

0.70

otal

0.70

iam

0.69

Activations Density 0.189%

references to various names, likely looking for specific individuals or entities

prefixes and suffixes related to names and places

No Comments

No Known Activations