INDEX

Explanations

the word "cluster" at varying levels of intensity

oai_token-act-pair · gpt-3.5-turbo

references to "clusters" 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 @ 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

hran

-0.80

estamp

-0.73

Ö¼

-0.73

inburgh

-0.69

 Dame

-0.69

toc

-0.67

issance

-0.67

esty

-0.67

ODUCT

-0.66

PLIED

-0.65

POSITIVE LOGITS

fuck

1.07

bom

0.97

 cluster

0.91

 clusters

0.91

usters

0.81

mates

0.79

 clustered

0.69

 munitions

0.68

mun

0.67

 grouping

0.66

Activations Density 0.017%

the word "cluster" at varying levels of intensity

references to "clusters" in various contexts

No Comments

No Known Activations