INDEX

Explanations

words related to co-working spaces and interactions with coworkers

oai_token-act-pair · gpt-3.5-turbo

references to workplace dynamics and relationships

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

anism

-0.84

icons

-0.78

anamo

-0.76

ulic

-0.74

hene

-0.73

tarians

-0.71

alos

-0.71

itol

-0.71

ogo

-0.70

ica

-0.69

POSITIVE LOGITS

 roomm

0.97

ãĤ´ãĥ³

0.86

 coworkers

0.83

yang

0.79

 roommate

0.76

zona

0.75

xes

0.74

 neighbors

0.73

ĸļ

0.72

 girlfriends

0.72

Activations Density 0.043%

words related to co-working spaces and interactions with coworkers

references to workplace dynamics and relationships

No Comments

No Known Activations